The autumn movements of the Woodpigeon

Transcription

1 Bird Study ISSN: (Print) (Online) Journal homepage: The autumn movements of the Woodpigeon R. K. Murton & M. G. Ridpath To cite this article: R. K. Murton & M. G. Ridpath (1962) The autumn movements of the Woodpigeon, Bird Study, 9:1, 7-41, DOI: / To link to this article: Published online: 18 Jun Submit your article to this journal Article views: 474 View related articles Citing articles: 10 View citing articles Full Terms & Conditions of access and use can be found at

2 The autumn movements of the Woodpigeon By R. K. Murton and M. G. Ridpath* (Ministry of Agriculture, Fisheries and Food, Infestation Control Laboratory, Field Research Station, Worplesdon, Surrey) Received 16 July 1961 (A Report to the British Trust for Ornithology) INTRODUCTION THE W O O D P I G E O N (Columba palumbus) in Britain is described by Witherby et al. (1940) as being both an almost completely sedentary resident and also a winter visitor from abroad. The winter immigrants were said to arrive on the east coast, between the Thames and the Northern Isles, from September to January but chiefly in November; and their numbers are stated to be small in some years but `immense' roughly every four to six years. Alexander (1940) believed that the winter flocks in southern Scotland and northern England came from the Highlands and that some autumn flocks on the East Anglian coast came from Scandinavia. He quotes Pennant (1776) as bringing forward the view that some of the British winter population bred in Sweden; on the other hand, later writers such as Montague (1802) and Selby (1835) considered that the winter flocks could be accounted for entirely from the population breeding in Britain. Colquhoun (1951) pointed out that the so-called `foreign' birds were distinguished by plumage characteristics which were really those of juveniles and thus were no proof of foreign origin in themselves and he considered that there was not enough evidence to show that the East Anglian winter flocks were indeed immigrants. Alder 0949, 195,), Wilkinson (195o), and Lack (1954) described autumn flights of Woodpigeons on the coast of Sussex and Kent which Wilkinson considered to be arrivals but this was disputed by Lack and Ridpath (1955), who showed that the apparent arrivals on the south coast could be explained by an observed tendency for pigeons to fly a short distance out to sea over the Channel and then return to the coast, giving the appearance of arriving migrants. The origin of these southerly flights was traced to inland roosts, which appeared to build up in numbers suddenly at the end of October and beginning of November. As regards movements within Britain, Colquhoun (1951) said that there is regular migration from the north to the south of Britain, both in autumn and mid-winter, much of it along a well established route of 400 miles down the east coast. This contention was based mainly on winter and summer flock counts in England and Scotland. Beeston *Present address Wildlife Survey Section, C.S.I.R.O., Hobart, Tasmania. 7

3 BIRD STUDY (1931) recorded a south-westerly movement along the Stour valley in Worcestershire over a number of years from October to December. Lack and Ridpath (1955), from their observations of roost movements in Oxfordshire and Kent, wondered whether an apparently southerly bias in morning flights from roosts, which built up rather suddenly in autumn, was connected in some way with a southward migration. Ash, Ridley and Ridley (1956), commenting on this, suggested that they might be feeding flights or local movements. Thus it can be seen that many conflicting ideas about the origin of autumn and winter flocks of Woodpigeons have been put forward. One thing, however, which is clear is that flocks containing large numbers of Woodpigeons begin to be seen from the end of October onwards, both inland and at the coast. The problem is whether these flocks are of local or distant origin. This may be resolved into two questions : I. Do any Woodpigeons migrate to Britain from abroad? 2. Do any Woodpigeons migrate within or from Britain? Two lines of approach existed direct observations and ringing. As far as ringing is concerned it is possible to examine the recovery data of the B.T.O. ringing scheme, and an analysis of these is given first in this paper. The other method, that of direct observation of movements, requires a series of watches from one position. In view of the confusing observations and explanations already mentioned, it seemed worth while trying to use this method on a large scale, in order to see whether or not there was a preference for any particular compass direction in Woodpigeon flights in early morning in autumn. Accordingly the B.T.O. accepted this enquiry as a Trust-aided investigation in 1955, with the authors as its organisers. Observations were made by members of the Trust and also by certain regional officers of the Ministry of Agriculture, Fisheries and Food. BRITISH RINGING RECOVERIES About one hundred recoveries of ringed Woodpigeons were available for analysis when Colquhoun's (1951) Woodpigeon Report was written. Today, about five times this number have been recovered so that a more extensive analysis can be attempted. Only those results which have a direct bearing on migration have been included in this paper, but life table analysis and age breakdown studies can also be made from the available material and have been presented separately (Murton, 1961). If there are any grounds for believing British Woodpigeons to be migratory on a large scale then the tendency should be apparent in the total of recovered marked birds. In Table I the distances moved 8

4 MOVEMENTS OF W'OODPIGEONS by birds from the place of ringing are given. Only those individuals ringed as nestlings and recovered during the period November to March are considered, so that any migratory tendency present would not be obscured by summer recoveries. The age on recovery being known, it is possible to separate the proportion of first year pigeons from adults in this sample and to test for any differences in movements found between them. It is clear from the table that the majority of Woodpigeons are sedentary, the distances covered by most individuals TABLE I-DISTANCE MOVED BY WOODPIGEONS PROM PLACE OF RINGING (ringed as nestlings, recovered November March inclusive) Distance moved in miles Age in months on recovery Io Over I II II Total Percentage of first year birds Average age in months.. 29.o Note: The differences in the age proportion of young recovered in each distance group are statistically significant, x2 =26.2; d. =3; P< ooi. being no more than would be expected from local feeding flights. In fact it is striking that in a species noted for its powers of flight,, 46 per cent of the birds travelled less than five miles, and 75 per cent of the total moved less than twenty-five miles. The most distant recovery was of a bird ringed at Blagdon, Northumberland in August, 1944 and recovered at Tipperary, Ireland, 15 months later representing a flight of 330 miles SW. An increasing percentage of young birds is found among those birds moving greater distances. In fact, of those moving more than twenty-five miles (and up to 33o miles), 78 per cent were in their first year of life. The direction and distance covered was determined for io8 birds, which moved more than 5 miles. Those which were less than 12 months old were separated and the results for the two groups are presented in Table II. Both groups show the same flight preferences and so the results were added for statistical analysis. Although not significant, the figures suggest a preference for flights to take place in southerly directions and a reluctance for birds to fly in westerly directions. The numbers moving in other directions do not differ from expectation. The average distance moved, although not high in any direction, tends to vary in the same way. The ringing recoveries indicate that most Woodpigeons are sedentary but that juveniles make the longer flights and disperse from the place of breeding. Furthermore this dispersal may have a southerly bias. Adults if they do move, seem also to prefer southerly flights and a proportion (c. 25 per cent) of the population may make flights of over 25 miles. 9

5 BIRD STUDY N N U F ^ v^ wx ^ ^ N ri) M ^D O N ^ n r/d N W N Q ^C7 " O > 3 ^ b N I N N z á ^ w 2 N W N J W ^ Z ón o ^ 2 F... á F H ó z I O

6 MOVEMENTS OF WOODPIGEONS A test was made to determine whether the geographical locality of ringing affected the tendency for birds to move. All recoveries (adults and juveniles) were grouped according to the area where ringed and the numbers moving more than five miles were separated. The results are given in Table III. A varying percentage of birds moving more TABLE III-DISTANCE MOVED BY WOODPIGEONS RINGED IN DIFFERENT AREAS Number Percentage recovered recovered Average distance moved in miles Total more than more than of birds recovered at Area recovered 5 miles,5 miles over 5 miles Scotland 151 3o NE England NW England SW England SE England East Anglia Total Note: The difference in the number moving more than five miles in the different areas is statistically significant, X 2 =24.3, d.f. =5, P< The areas selected for comparison comprised the following counties: NE England; Northumberland, Durham, Yorkshire, Derbyshire, Nottinghamshire, Lincolnshire, Leicestershire, Rutland, and Northamptonshire. NW England; all counties from Cumberland to Warwickshire, west of the above, and north of Gloucestershire, including Wales. SW England; Gloucestershire and Wiltshire and all counties west. SE England; Oxfordshire, Berkshire, Hampshire and counties to the east, excluding those in East Anglia which were Essex, Hertfordshire, Bedfordshire, Huntingdonshire, Isle of Ely and Norfolk and all counties east of these. than five miles occurred in different areas. The lowest percentage of moving birds was recorded for those individuals ringed in Scotland; which does not support Colquhoun's idea that Scottish Woodpigeons migrate to the south and are responsible for increases in flock size in southern England in the winter months. An analysis of the direction taken by the birds moving more than 5 miles did not show any marked differences between areas and the general picture was the same as that given in Table II. Much of the variation may have been due to local differences in environment. Thus the high proportion of birds moving in NW England is affected by the totals for the Lake District and it is possible that there is more tendency for birds to leave the woodland areas in winter to feed on the surrounding arable land. This view is supported by field observation in the Ormskirk area of Lancashire. This area has little nesting cover and a low breeding population, but the food supply is good. In winter large flocks of Woodpigeons occur, which are certainly not from the summer population and which many local farmers believe come from further north in the country. Without more information it is not possible to discuss such local variations, which are probably governed by many different factors. There are also variations in the average distance travelled by birds from different II

7 BIRD STUDY areas but considered in relation to the proportion moving more than 5 miles they do not indicate any marked population shifts. The results do show that large movements from north to south in Britain do not occur. If Woodpigeons first tend to move in November it might be apparent from the ringing results. Recoveries should be mainly local during the breeding season but show an increase in the distance travelled in November or at whatever time movements occur. In Table IV the percentage of birds moving more than five miles is given for each month for nestlings ringed from July-September inclusive. From the table it is seen that in September and October most recoveries TABLE IV-DISTANCE TRAVELLED EACH MONTH BY NESTLINGS RINGED JULY-SEPTEMBER INCLUSIVE Month of recovery Total recoveries.... No. recovered more than 5 miles from place of ringing Percentage recovered more than 5 miles from place of ringing *Average distance in miles recovered from place of ringing Sept.- Oct. Nov.- Dec. Jan.- Feb. Mar.- April May- June Notes: *The average distance moved from place of ringing refers only to those birds moving more than 5 miles: local recoveries are not included in calculating the average. The differences in the proportions moving more than 5 miles in each bi-monthly period are statistically significant, X 2 =53.2, d.f. =5, P<o.00t. are from near where the birds are ringed but that in November and December there is a marked increase in distant recoveries, which continues into January and February. The proportion recovered away from the place of ringing then falls between March and August suggesting that birds, which have left their place of origin, tend to return to the same area in the following breeding season. The figures also suggest that some Woodpigeon movements might occur during the early spring, associated with a return to the breeding places. However, these movements affect only a small proportion of the population and these are mainly juvenile and first-winter birds. In only 15 out of 492 recoveries had the birds moved more than too miles. There are a few records involving a sea crossing, mostly from the Isle of Man to the mainland. There are four recoveries* of British *Since this article was prepared, there have been three further recoveries abroad of British ringed Wood-pigeons. They are as follows : Ringed young. Friskney, Lincs. recovered mid r.6o Gourin, Morbihan, France Ringed young. Elmers End, Kent. recovered it.it.6t Riaillé, Loire Atlantique, France Ringed young. Weston Colville, Cambs. recovered Nr. Marquise, Pas de Calais, France. July- Aug I2

8 MOVEMENTS OF WOODPIGEONS ringed Woodpigeons on the Continent. It is of interest that all occur between November and December. They are as follows: RS 2714 Ringed yound Lower Culham, Oxon. recovered St. Calais, France Ringed young Cuckham, Berks recovered Finistere, France Ringed young Sidlesham, Sussex recovered St. Malon-sur-Mer, Ille et Vilaine, France Ringed young Sidlesham, Sussex recovered St. Ouen La Roverie, I11e et Vilaine, France Only one foreign Woodpigeon has been recovered in Britain. It was ringed as a nestling in Denmark and was recovered 8 years 8 months later in Queen's County, Ireland. WOODPIGEON MIGRATION IN CONTINENTAL EUROPE In Europe a well marked migration of Woodpigeons takes place each autumn from NE to SW, which it is interesting to compare with the situation in Britain. It is documented both by ringing recoveries and field observation. We have made a search of the literature on this subject and give a summary of the results below. Much of it has been found in the annual ringing reports of various countries in Europe, many of which give only the numbers of birds recovered, omitting the numbers ringed to date. Another limitation of such data, from our point of view, is that they are often selective, only containing those recoveries likely to be of special interest because of the great distances travelled by the birds or because the records show an exceptional age for the species. Nevertheless these drawbacks detract only slightly from a clear overall picture. Of the Scandinavian countries, Swedish workers have ringed the highest total of Woodpigeons. Details are given in the annual ringing reports of the Goteborg Natural History Museum (Fontaine, ) A total of 541 Woodpigeons had been ringed up to and including 1955 from which there were 43 recoveries. Of the recovered birds, about 29 had made long flights; 23 were recovered in France (mainly in the SW including many in the Gironde Province, four in the Pyrénées), two in Spain, two in Germany and one each in Denmark and Belgium. In Norway the total Woodpigeons ringed up until 1953 appears not to have exceeded 200 birds. (Olstad, 1954; Holgersen, 1954.) We can find only two published accounts of recoveries for these birds in the Norwegian literature, one being recovered in Norway and the other in the Pyrénées. In Finland 6o W.00dpigeons had been ringed up to B 13

9 BIRD STUDY 1953 and of these, five have been recovered further south in Europe (Valikangas & Nordstrom, 1954). Woodpigeons do not breed in large areas of Scandinavia. Witherby et al. (1940) give the northern limits of the species as 65 N in Norway and Sweden and 64 N in Finland In Finland, Merikallio (1958) says that it is clearly a southern species and occurs chiefly in the SW corner of the country rapidly decreasing to the north and centre. Using a line transect method, from 1941 to 1956 he estimated the Finnish population to consist of 164,000 pairs. For a species whose favoured habitat consists of cultivated land with woodland edges even some of that part of Sweden lying inside the pigeon's northern limit is also quite unsuitable, consisting of mountains and extensive coniferous forests, where pigeon density is low except at those edges which adjoin cultivated areas. Probably for the same reason, Vik (1953) states that the Woodpigeon is abundant only in the low-lying area of SE Norway and in the area around Trondheim, which forms only a small part of Norway. Observations by one of us (M.G.R.) in Jutland and southern, western (to 62 N) and south-eastern Norway in June and July 1956; and in southern Sweden and in parts of southern, central and northern Finland, and northern, central and western Norway in June and July, 1957 confirmed the lack of Woodpigeons in northern Scandinavia. This suggested that in the rest of Scandinavia they were not only largely restricted to cultivated areas and their edges but were also less abundant in such areas than in eastern and southern England. Scandinavian and Finnish workers say that Woodpigeons leave their countries before the winter months. The route taken by Scandinavian and other north European Wood - pigeons appears to take them to the west coast of France from the Gironde to Biarritz. This is not only shown by ringing recoveries but also by direct observations such as those of d'arne (in Mayaud, 1940) and Owen and Burton (1953). On arrival at the Pyrénées, many pigeons cross into Spain mainly at the low-lying western end, although some do cross the much higher mountains near the centre of the range (Lack, 1953). None appear to cross at the eastern end of the Pyrénées. Barriety ( 1956, 1957) using the traditional method of mass trapping described by Lack (1953) has caught and ringed large numbers of pigeons during their autumn passage through the western Pyrénées. Three of these birds have been recovered in Spain and Portugal during the winter months and others have been recovered in north Europe in spring and summer during and after their return NE. One bird ringed in the Pyrénées in October 1954 was recovered in November 1955 in Sicily. From inclusive a total of 1,538 pigeons was ringed producing 125 recoveries. A further total of eight birds marked elsewhere was captured during the four years. Recoveries at the 14

10 MOVEMENTS OF WOODPIGEONS FIGURE I. Map of Europe showing main recorded autumn passage of Woodpigeons. Filled circles indicate winter recoveries in W. France of four British ringed birds. The arrows for Continental movements are based on visual observations, supported by ringing recoveries; the British movements on visual observation only. The dotted arrow represents the tendency for British birds to fly south in the autumn.

11 BIRD STUDY station of birds marked elsewhere include a Woodpigeon which was ringed in the nest in July 1952 in Finland and was caught in October The observations of all the authors mentioned above who have rerecorded the passage of Woodpigeons through SW France and the Pyrénées agree that October is the main period for migration with a peak between the second and third week. By the end of the month the migration appears practically to cease. A similar picture emerges from observations between August and September from at the Col de Bretolet and Col de Cou in the Swiss Alps by Godel & Grousaz (1958) and Vuilleumier (1959) The mountains force the pigeons to take a SW direction at this point therefore their undeviated standard direction is hard to determine precisely, though it must be within the southern arc. The earliest date of passage was 15 September and the latest date 26 October. The biggest flocks seen were 564 on r1 October 1958 and 300 on i October It is possible that some of these pigeons finally reach Italy where Toschi (1950) records an annual migration in autumn into central Italy, whose standard direction is SW. A returning migration in spring through Switzerland has been recorded by Geroudet (1958) from mid-february with a peak in mid-march and ending at the beginning of April. The observed movements of Woodpigeons in Europe are summarised in Fig. 1. EXTENSIVE ENQUIRY, Potential observers were sent record cards and a sheet with instructions and information about the enquiry. Each observer was asked to choose a watching point with an extensive view, but where local features would not be expected to cause a deflection of the birds' flight. They were asked to record the date, place, times of beginning and end of watch and the total number of Woodpigeons seen flying in each of the eight compass directions between dawn and moo hours each day, from October to December Similarly numbers and directions of any seen returning to the roost in the evenings during the last two hours of daylight were asked for, in an effort to see whether morning movements could be traced back to a roost in the evening. We also asked for any directional counts at the roost of birds leaving in the early morning period. Observers were requested to record all negative watches. Finally, we asked for counts of any Woodpigeon flocks seen flying to land from the sea at any time of day. Over 2,000 record cards were distributed and 147 observers sent us records in response. A total of 117,857 individual flights distributed between the eight cardinal compass points were recorded from all areas of the British Isles between 9 October and 24 December Observations before 1 6

12 MOVEMENTS OF WOODPIGEONS ó ^o.^oz N ^n o0 ^n ;, ^ v Np a o, o ^ z ḏ ó ^.^ S o0.71 0o N ON h 00 0 N `O s ^ o N z h ^ á.. ^ 0o O o ^ b H ^ ^ oo,0 en 00 '+ h N 0, 'ce`n b 7 00 CS 0; 00 v+ h N N i.v O 7ON NN MN.} M ^ y 00 n 0 zm ^ ^ b O N ~ `O a h b ON ON ON á N0N0 r^n MOO 0 n h N a h 0 00 h O r r Ó ^ p ^ ^ N h b M V M W ^.hi r ḥ i o0 0 vi N ON Ó N^ h O á ^t N q N.+ nn nn N 00 ^ N I v;^p 00 h M h M SO O^. O^ 00 é+ h ^ M U h hoo hv;,511 s9:1-^ CO'd' O.. r; N.+.i rn ry 00 P p o0o á m 01 ^NO 0 CO M Ó T ^NO N ^O W r ^ N h oo h o0 '; e -., ^.n ^ O LL M M t+; N...., 00 ^ N^O ON V N O, ^ h o0 N 0 ^D,n.. O vi,n.+ N ^O h N N...+ N Os rn 00 [.14 os 1-1 h.1. N 0, 00.+ N `O b n N z N NON v-1 O 00 a rn N O.} h N N,d N N0 O.00 vi b.^.^.n v-1 z p n p,o m N N a N 0` b O h7 N N a^ ó ^M b ^ ^ ^ nn N rn.y N N.i.' 0 ^ V ^ V 0v O [3, o z ^ 17

13 BIRD STUDY FIGURE 2. Direction taken by Woodpigeons on leaving their roosts in the early morning during the period 9 October-24 December, and after these dates were too few in number to he of value. A weekly analysis of these flights is given in Table V. These data were treated statistically (using an analysis of variance, the figures being transformed into the number of pigeons seen per observer hour and the square root of these figures considered). The variance ratio for directions against residual showed significant differences (P<o.ot) in preferred flight direction. Most birds flew in a NE to S direction and relatively few to the N, NW and W. The wide scatter, from NE to SW might have been reduced if the period of observation had been restricted to dawn till o800 hours, instead of i000 hours, thus eliminating later, more random feeding flights (Fig. 2). 1 8

14 MOVEMENTS OF WOODPIGEONS Table V shows that there was an increase in the total number of birds seen per observer hour in November, and this was significant (P<o.00,). Two sets of data are of particular interest because they can be related to other observations in SE Kent. G. C. Phillips watched for ten mornings and eleven evenings between 16 October and 3 1 December 1955 at a roost at High Halstow, north Kent on the south bank of the Thames. The results of his counts are given in Table VI. It TABLE VI NUMBER OF WOODPIGEONS USING A ROOST IN NORTH KENT IN 1955 (observations made daybreak-183o hours and hours dusk.) Direction Date N NW W SW S SE E NE Total 16 Oct. in o Oct. out in Io o Oct. out o o in a Nov.$ out o o I in o loo Nov. out o o o in o o NOV. Out NO OBSERVATIONS in o t 20 Nov. ** out o o in NO OBSERVATIONS 27 Nov. out o 0 7 I in o o Dec. out o 0 o o 8o in o $ II Dec. out o o o * in o i8 Dec. out o I in o Dec.,out NO OBSERVATIONS in o Dec. out o 0 0 o I TOTALS out S in Notes: Data supplied by G. C. Phillips. Morning departures are labelled `out', afternoon arrivals 'in'. The directions given in italics for afternoon arrivals refer to the direction from which the birds flew, they are thus complementary. *Rain and poor visibility affected counts. 't Shooting took place in the wood. t6 Nov. At 0815 hours 25 birds flew into roost from NE may have crossed River Thames. 12 Nov. At 0751 hours Ito appeared from N flying high, probably crossed river. Circled roost and joined birds leaving E. Included in above total. **20 Nov. At 0825 hours 3 birds flew into roost from N. can be seen that an increase in morning departures occurred around 12 November. His previous morning watch was carried out on 6 November when rather small numbers were involved, but a vastly greater total returned to roost in the evening. A similar increase 19

15 BIRD STUDY occurred in SE Kent on 9-to November at a roost watched by one of us (R.K.M.) (see below). Phillips observed Woodpigeons flying south across the Thames during the early morning watches (see notes accompanying Table VI). A decrease in roosting numbers occurred at the end of November but there seems to have been another increase in late December. In SE Kent a second increase was also noted but it occurred on 6 December. Phillips points out that there was a conflict between east and south and that the birds which flew east eventually turned to fly south in order to cross a ridge of high land FIGURE 3. Direction of flight of Woodpigeons seen during the early morning near Leith Hill, Surrey, for the period z8 October I2 November,

16 MOVEMENTS OF WOODPIGEONS further from the roost. Thus local topography, to some extent, affected the flight direction, but southerly points were clearly preferred during November. Returns to the roost took place from the same directions as the morning departures with the exception that morning departures southwards in late November were not balanced by returns from this direction. Phillips believes that birds returning from the south in the evening could not see the roost which was hidden by a ridge, therefore they flew north until reaching the Thames where they turned to arrive from the east which was the preferred departure direction. One of us (M.G.R.) watched between dawn and t000 hours, from 28 October until 12 November 1955 on the Tower at Leith Hill, Surrey (965 ft. O.D.) and from a point 4 miles NW on Hackhurst Downs (70o ft. O.D.) from 13 November till 16 November. The Leith Hill Tower is in a commanding position overlooking the Weald and has an excellent view in all directions; the view from Hackhurst Downs was not quite so good. Fog made observations impossible from 3 to 6, 8 and 16 November and no watch was kept on to November. A total of 4,359 flights were seen and the distribution of compass directions is shown in Fig. 3, which shows a clear preference for flying south with some going SW. The broad picture that emerges from the extensive investigation is of an increase in the number of pigeon flights in early November and a preference for flight directions between NE and S. This was most clearly shown from mid-november onwards until the investigation ended on 24 December but it was also in evidence from 9 October, though less clear-cut. We received no definite records of immigration. We made a second appeal for information in the autumn of 1956 but only 12 cards were completed and the data were consequently inadequate. OBSERVATIONS IN SOUTH-EAST KENT What was the origin of these often spectacular morning flights and how exactly did the flocks build up? To answer such questions we decided to study one area intensively in autumn. The area chosen was in SE Kent, within a five mile radius of the South Foreland, which was already known well because we had made observations there between 20 October and 15 November 1953 and between 1 November and I December 1954 (Lack & Ridpath, 1955). The countryside is open and characterised by a series of long, low, parallel ridges running NE to SW. Much of it is good agricultural land under crops and there is some pasture. Small deciduous woods are scattered evenly across it and there are also a few coniferous plantations. The general level of the land is 200 feet O.D., with a maximum of slightly over 40o feet, and the area is bounded by an abrupt coastline formed by high chalk cliffs. To recapitulate briefly our observations there in 1953 and 1954, we 21

17 BIRD STUDY had found that large flights of Woodpigeons (maximum of 2,27o in a morning) occurred between dawn and moo hours, mainly before o800 hours, with a sudden increase on to November 1953 and 8 November Observations at the coast showed that the pigeons arrived from inland, i.e. from the northerly quarter. The coast deflected these flights when they reached the cliffs causing pigeons to fly back inland, to settle or to coast westwards, and only a few continued out to sea, between S and SE; but were always seen to turn back to the land. We had traced some of these flights back to a roost 5 miles inland at Waldershare, to which the birds returned in the evenings. Observations from the South Goodwin light vessel by one of us (R.K.M.) from I November until she was wrecked on 27 November showed that no Woodpigeon flights could be seen at sea 4 miles due east of St. Margaret's Bay. The roost at Waldershare was in a mixed deciduous wood of about 25 acres in Waldershare Park. Another roost was situated 12 miles due north in a mixed deciduous wood of about 70 acres near Dane Court at Tilmanstone. Both woods were on sheltered slopes and were generally undisturbed as they were used for pheasant rearing. In the autumn of 1955 one of us (R.K.M.) watched the Waldershare roost from 29 October to ii November from just before daybreak ( hours) until morning flights had finished, between 0830 and 0900 hours. The number of Woodpigeons leaving the roost each morning and also the number arriving during the same period is given in Table VII. After daybreak some birds could be seen to leave the roost, fly round and return again. There was also a certain amount of flapping about among the tree-tops with short flights within the roost. Later in the morning some birds returned to the roost having already spent some time feeding. This is typical of the summer months and becomes less noticeable when most of the day is required for feeding. The totals in Table VII exclude the abortive sallies already mentioned and other arrivals, so that the figures represent the overnight totals of birds in the roost. It can be seen that numbers rose suddenly on 9 and It) November and this date of increase is remarkably close to the increase in numbers noted at the coast in 1953 and 5954 (10 and 8 November respectively), and in north Kent in 1955 (about 12 November). Associated with the increase in numbers there was a noticeable change in flight behaviour. As can be seen from Table VII, the birds left in many small groups up to 8 November but on. so November the departures were in six large flocks. Until 8 November, on leaving the roost the birds flew mainly to points around NE with smaller numbers flying SE and S and very few moving in other directions. Many flocks settled to feed within sight of the observation point. However on so November, nearly all the birds flew towards the.se bringing them directly to the coast. Not only were they in large flocks but their flight was more direct and 22

18 MOVEMENTS OF WOODPIGEONS all flew out of binocular range. The impression of purposeful flight in large flocks was in striking contrast to the earlier observations. It was known that Woodpigeons were also roosting in the woods at Tilmanstone and departing birds could be seen leaving these roosts, while watching at Waldershare. Mr. A. Hogbin, head keeper of the Dane Court estate, estimated numbers in the Tilmanstone area to be around 300 but on to November, a marked increase also occurred there. About 1,40o were seen to leave the area from the Waldershare watching point and he confirmed that a large increase had in fact taken place. TABLE VII NUMBER OF WOODPIGEONS USING WALDERSHARE ROOST IN 1955 (observations made between hours approx.) Total departures Total arrivals Date o600-o800 hours o600-o800 hours Birds Flocks Birds Flocks 29 Oct I 3o Oct * Oct. 259* I 1 Nov Nov. No watch kept 3 Nov Nov. 236* I 5 Nov. No watch kept 6 Nov... No watch kept 7 Nov Nov. 2o0* Nov. No watch kept io Nov. 1,446 6 it Nov Dec..... No watch kept 914t 7 Dec , Notes: *These totals do not include birds seen arriving, and which subsequently left. They therefore represent overnight roosting numbers as do all others. t Total refers to afternoon watch of number returning hours. From 29 October to 7 November Woodpigeons also entered the roost during the early morning, but in nearly all instances they were the birds which had been making abortive sallies. On 8 November birds entered the roost from further inland, i.e. from the west and 6o arrived in one flock in this way, followed by 20 which separated themselves from a flock of 54 to enter the roost, while the rest of this flock passesd over. In addition, on this morning 139 other birds passed over the roost from further inland, a situation not noted previously and 23

19 BIRD STUDY then 27 Woodpigeons dropped almost vertically from the sky from beyond eye range into the roost. On io November, 165 birds passed over the roost flying east during the morning watch. On 9 November no morning watch was kept and instead, time was devoted to counting birds returning to roost in the Tilmanstone district but nothing was seen which would have accounted for the marked increase at Waldershare on the morning of 10 November. Thus observations in Kent in 1955 showed that the increasc noted at the coast in the two previous years in early November could be accounted for by an increase at inland roosts at the same time. The other evidence mentioned above indicates that this inland build-up of numbers was not confined to the one roost under observation but was a general increase in numbers in the locality. The origin of the increase was not known but from impressions already gained of local population numbers, it is thought unlikely to have been from local sources. In this connection the observations made by G. C. Phillips in north Kent are particularly interesting. His roost records (Table VI) suggest that there was an increase in numbers from the north with a loss to the south which suggested a north-south movement. Furthermore, his roost increase occurred at the same time or possibly just before the increase at Waldershare. In the autumn of 5957, a team of three watched the Waldershare roost from dawn until 0830 hours and from 1300 hours till dusk from three vantage points, almost equidistant round the roost and each about 300 yards from it, thus ensuring that no movements out of or into it would be missed; and during the rest of the day one observer remained at the roost to note any arrivals or departures. The results are given in Table VIII. The morning departures were usually over by 0830 hours and no birds entered the roost before mid-day. For this reason the morning totals have been labelled `out'. The return to the roost began with small numbers shortly after 1200 hours and went on till dusk, therefore the evening totals have been labelled `in'. The distribution of morning flights between the eight compass directions showed a marked preference for SE and S as compared with other directions. Very few flights occurred between NE and NW. Table IX shows that the number of pigeons leaving the roost each morning between 29 October and 4 November was greater on every day than on any of the mornings from 5 to 13 November. This decrease was preceded by a peak of 837 pigeons on 3 November. Column 3 of Table IX shows that from 1 to 3 November more pigeons returned to the Waldershare roost in the afternoons than had left it in the mornings of the same day. From 4 November onwards fewer birds came back in the afternoons than had left in the mornings, except on 5 November. For the directions in the arc NE, N, NW, W, the numbers coming to the roost in the afternoons were usually greater than the numbers leaving it in the mornings in these 24

20 MOVEMENTS OF WOODPIGEONS TABLE VIII NUMBERS AND DIRECTIONS OF WOODPIGEONS LEAVING AND ENTERING A ROOST AT WALDERSHARE 29 OCT. TO 13 NOV., AND 16 NOV., 1957 (top figure for each date represents number out in the direction shown, bottom figure the number arriving from this direction) ( 1 ) (2) (3) (4) (5) (6) (7) (8) (9) (Io) Date N NW W SW S SE E NE Direction of flight not seen Left roost late afternoon 29 Oct. out in o 7 io NO WATCH 36 o o 3o Oct. out o 2 46 to o itl NO WATCH 31 Oct. out o in NO WATCH 1 Nov. out o o 2 Nov. in o out o Io in I Nov. out in o Nov. out o ill Nov. out o in o Nov. out o Nov. in out o I I in o o 8 Nov. out o 0 o I in Nov. out in o To Nov. out in ii Nov. out o I ill Nov. out o in NO WATCH 13 Nov. out o 0 o Io in NO WATCH 16 Nov.* out it NO WATCH TOTALS Out I-II Nov. in o *Two observers only. directions throughout the period of observation, departures exceeding arrivals on only 8 out of 44 occasions; whereas, for the directions in the opposite arc this was only true until 3 November and after that date departures to the E, SE, S and SW generally exceeded returns from these directions (the opposite occurring respectively in exactly a quarter of the occasions in both cases). The 4 and 5 November were days of change during which Waldershare passed from the period of afternoon gains (r to 3 November) to the period of afternoon losses (6 to II November). In other words, the Waldershare roost gained pigeons from 25

21 BIRD STUDY all directions until 3 November after which it lost them to the easterly and southerly quarters while continuing to gain them from the northerly and westerly quarters. If the number of pigeons which left the roost in late afternoons is subtracted from the number which entered it, the resulting totals (col. 5 of Table IX) on the whole show even greater differences from the morning totals than those in column 3. Except on 2 November the number of birds counted into the roost and still known to be there at dusk was substantially smaller than the numbers seen to fly out of the roost in the morning of the same day. If we consider the totals for the entire is days, we find that for 3,244 morning departures there occurred 4,073 returns to the roost but that the known remainders at dusk only came to 1,699, because of the 2,374 departures during the late afternoon (Table VIII col. so). These birds appeared to fly off to the north, where we suspect that they joined the roost at Tilmanstone, and also to the west where another roost existed from the direction of which we often saw pigeons passing over in the mornings. If the number of birds counted into the roost and still known to be there at dusk is compared with the total seen to leave the roost the following morning, the figures in Table IX col. 6 are obtained. They show, with one exception, that many more birds left the roost each morning than were known to enter the roost and stay there the previous evening. The differences between the counts of returning pigeons in the afternoon still known to be in the roost at dusk and both the counts of the departures on the same and the following mornings have three possible explanations. Firstly they may be due to a failure to observe and count all the returning birds, secondly, to the existence of a larger number of pigeons in the roost than left it on any one morning (i.e. many remaining in the roost all day) and, thirdly, to arrivals of pigeons at the roost in the dark, when no observations were made. Although failure of observation must be responsible for some of the difference, we feel that it is unlikely to have been responsible for the fact that most morning totals differ by more than 5o per cent from the number counted into the roost and known to be there the previous evening. We saw pigeons moving about in the roost occasionally during the morning and early afternoon when few or none had returned from outside and we also disturbed them when entering the roost but it was never possible to make an accurate count of them. The presence of these birds may partially account for the fact that more pigeons could be seen to depart in the mornings than were seen to return and remain there at dusk on the evening of the previous day. However, for this to explain the situation there would have had to have been at least 1,438 extra pigeons remaining all day in the roost on r November, in addition to the 249 which did fly out that morning; and we are convinced that this was not so for we would have detected 26

22 MOVEMENTS OF WOODPIGEONS vc *1' _ 3... Q p.-, ^ ^.s^" I I I^N8M^4^P^á o I I I I yy;..li' C -9.(-) ++++ i + -F h^ q ittu '3 : v o F 6 F rn E^^ ^ R v.^ a ^ U I n I ^ a`' y M N^^^`^ O02 I I I I n^ii3 ^ 6 F.' ^+ ^ ^ ^ ^ ^ ^ ^ ^ ^ ^ ^ Q ^ +r., C9 F q Ue z z á w w w F í3 z F w wo N L1 w e S e,1 ^ Z m = ^ ti 6 w 4'1:0 U a 0.a- so co t-.. N N h H U' ^r N 1-S. s^ I I 1 o ^. 00 n ` en en,0,o,n v, NO +,0 I I I I v V PA C ^ rr, f"' : w ^UY AO CrQ 6 w W ''''''6' a Fa Uf z o z ^ ^ r ^.. '4' H s N C^ H ^ O ^O O en N N e0 V om ^,,, ^ vó I I I ^^l^mv P. S1 :,. i I I I w. ^ +++i+iiiiii OF ^i;ó '40 'CI ^i `-- ''' a a o q V) a a a O w -ír3 y-e., ^,6 ^ ^ C" s. ^ 33 3 benann ^ o O' N` 33 3 I N < á ^ ~ ^zzz - zzz Q w I Q a x 6.. a IA a al. O 0^ ^O r e0 en N.} r O a N `0 t `,._. c S 4 N l^ en d' l^ m D\ t` ti H b h M h U íni e. `^i F f--.o Q }d' rti tm h t'+ H N N H.a N h M > > > > > > > > > > > > > > I > > > U U V O 1- O^ I- O^ 1-^O I^O I- O^ 1-^ O^ O O O O U F O^ I-O^ 1-^Or 000Z ti ZZZZZZZZZZZZ ^OZ ^ZZ Q N eon en ^ N en d' 41,0 S W 010 H.Ni M^0 H

23 BIRD STUDY such large numbers had they been present in a small wood (25 acres). As regards arrivals at the roost in the dark there is little evidence available. We know of only three records of Woodpigeons occurring at a lighthouse light at night. Two were reported by D. F. Billett (per J. S. Ash) at the lighthouse at St. Catherine's Point, Isle of Wight in 1957, one being seen at the light on the night of 30/31 March and two on the night of 5/6 April. The third record is from the Shipwash Lightvessel in the Thames Estuary where two pigeons circled the light at night on 25 March 1960 (Williamson, 1960). Williamson also records presumed night arrivals at Skokholm from March to May, Thus it is very possible that some night arrivals took place and there is also the possibility that some pigeons, which were in the area during the day did return to the roost at late dusk after we had left. Of the three possible causes for this excess, late dusk and possible night arrivals seem to be the most likely. Counts were also made of the pigeons leaving the roost near Tilmanstone and the results were similar to those at Waldershare, with a peak in numbers on 6 November followed by a decrease. Southerly flight directions were also preferred. Directional preferences at these two roosts agreed broadly with those for the extensive enquiry of 1955; SE flights from Waldershare and Tilmanstone roosts would take the pigeons to the coast at about South Foreland and St. Margaret's Bay, where we had in fact observed flights from the NW in the mornings in 1953 and In order to verify this, we posted an observer near South Foreland from 0725 to 0830 hours each morning from 8 to 17 November. The results of these watches are given in Table X. The numbers were very great (c. 16,000 on 12 November), much greater than could be accounted for by birds leaving Waldershare and Tilmanstone roosts, whose totals had already dropped to a low level by 8 November. In fact the observers at Waldershare and Tilmanstone had seen large SE flights passing across these roosts from further inland throughout the period of the watch, with a big increase on 5 November. When numbers at the coast reached the high totals of those on 12, 13 and 14 November, counting became difficult, particularly because most of these movements took place within the first half hour of observation. Two observers made the count on the 12th and three on the 13th and,4th and therefore it was possible for independent estimates of flocks to be checked against one another. As a further check on our ability to judge numbers, photographs were taken of certain flocks whose numbers had been estimated. These showed that our estimates were consistently lower than the real totals. The estimates for flocks up to too birds were close to the true values. A flock of c. 730 individuals was estimated by two observers to be 300 and 500, and similar discrepancies were found in all cases when such numbers were involved. On arrival at the coast the pigeons usually settled on stubble and 28

24 MOVEMENTS OF WOODPIGEONS ploughed fields within a half mile of the cliff edges but small numbers flew out to sea for a few hundred yards and then returned to the coast. Haze reduced visibility over the sea to about two miles. Little feeding appeared to take place among these settled birds and many rested on ploughed land, making no attempt to feed. By moo hours they had begun to drift back inland in small flocks. We tried to trace these movements back inland on six days (8, 9, II, 13, 14, 17 November) during the late morning and afternoon from observations at various points up to nine miles in an arc N to W of the South Foreland, but we saw no more than a tiny fraction of the coastal flocks returning. Yet, as in the case of the roosts, these huge numbers were seen again the following morning. Early morning observations were made on II November near Dover Castle, two miles W of the South Foreland, and on 12 November at Coldred, five miles NW, but the numbers of pigeons seen at both points were very small compared with the numbers at the coast, apparently arriving from this general direction, which were in the order of thousands (see Table X). TABLE X NUMBERS OF WOODPIGEONS ARRIVING AT SOUTH FORELAND IN (observations from hours) ate Number Nov. 7,400 9 Nov. 3, Nov. 1,900 II Nov. 4.40o 2 Nov. 16,500 3 Nov. 14, Nov. 11, Nov. Nil (fog) 16 Nov. No watch 17 Nov. 2,000 (fog) Notes: Flights were S to E but mainly SE on all days except for 17 Nov. when all birds were coasting SW. These concentrations at the coast gave every appearance of trying to depart southwards across the Channel and our impressions were of the marked similarity in behaviour to migrants such as the Starling attempting to leave a coastline but being held back by the unfavourable weather conditions (low cloud and haze or fog throughout the period). The sight of Woodpigeons in such numbers making short seaward flights over the water and returning to perch on the cliff edge in such numbers that several hundred could sometimes be seen perched on the actual cliff face, was impressive. Unfortunately the development of this situation was obscured by fog at South Foreland which lowered visibility to yds. between 15 and 17 November although observations at Waldershare on 16 November, where it was less foggy, showed a passage of 1,939 c 29

25 BIRD STUDY pigeons flying between E and SE across the roost from inland. No pigeons at all were seen at South Foreland on 15 November and about 2,000 were seen moving erratically SW along the coast in very poor visibility on 17 November, the last day of our observations. During November pigeons were seen to leave the coast at Portland Bill (see below). A small team of five observers was present in the area again in 1958 from I November to 5 December. Daily observations were made at three vantage points from dawn till 0900 hours. Neither in the early mornings, nor later were appreciable numbers of Woodpigeons seen in this area and both Waldershare and Tilmanstone roosts were practically empty. The Tilmanstone roost contained c. 300 birds until 16 November but this increased to c. I,000 on the 17th and 18th, dropping to c. zoo on the 19th. This seemed to be a small scale repetition of the peaks of previous years. The overall picture in the area was of a much reduced number of morning flights as compared with the same period in the preceding five years. In 1958 there was little evidence that numbers rose appreciably above the summer population and the situation was in marked contrast to that of The overall picture that emerges from these observations at Waldershare and Tilmanstone from 1955 to 1958 is of an increase in the numbers of pigeons at the roosts in early November, rising to a peak, followed by an abrupt drop. The nature of these changes is best understood by first considering what happened at Waldershare in November 1957 for which we have a full record for all the daylight hours. Column 1, Table IX shows that numbers in the roost increased from 1 to 3 November and were followed by a decline. It is significant that this increase was largely due to increases in the numbers of pigeons arriving at the roost in the afternoon from the S and SE, considerably in excess of the numbers which had left in those directions in the mornings on these dates (cols..5 & 6 of Table IX). What was the origin of these extra birds? The fact that departures from the roost in late afternoon were also much higher over the same dates (actually extending to 4 November) suggests that the increases, mainly from the S and SE, consisted of a number of pigeons whose choice of roost was labile. We suggest that this was because these pigeons were strangers to the area and its roosts. INLAND ROOST OBSERVATIONS A study is being made of Woodpigeon ecology by one of us in an area near Newmarket, Suffolk, during which observations made in 1957 and 1958 have a bearing on the increases in numbers of birds seen at roosts in November. The total population of birds inhabiting a large wood was estimated using the maximum occupied nest density as an indication of the number of pairs (Murton, 1958). Counts during the winter months of the roosting population showed that the total 30

26 MOVEMENTS OF WOODPIGEONS never exceeded that contributed by the nesting adults plus their young. However, an increase in numbers leaving in the early morning in November occurred due to two causes. Firstly, whereas before this time adults left the roost in small numbers throughout the day, they now departed in a few large flocks early in the morning. This was connected with changes in food availability and the total time required for feeding. Secondly juveniles were shown to be present in the roosts throughout the day in early October, and relatively few accompanied the adults to the feeding ground. In November a change occurred in their behaviour and they too left in the early morning with the adults. Behavioural changes were therefore responsible for what looked at first like an increase in population size. OBSERVATIONS ON THE EASTERN AND SOUTHERN COASTS OF ENGLAND, Apart from our own observations in SE Kent we have accumulated much evidence to show that an emigration of Woodpigeons occurs on the south coast in some years. Dr. J. S. Ash has supplied us with observations from Portland Bill for the autumns of In one Woodpigeon was seen on 16 October which flew S out to sea and returned N at 0835 hours. On 13 November one Woodpigeon flew SW and returned N. Watches were kept from dawn on 1-5, 13-18, October, 12 and 14 November, but these were the only pigeons seen. In 1956 ten morning watches were made from 2 October to 4 November and no Woodpigeons were seen. In contrast to these low totals of 1955 and 1956, on 2 November 1957, 750 Woodpigeons flew out to sea SW from the point and were lost to sight still flying SW, a direction which would have ultimately taken them to Britanny.. In 1958 again only small numbers were seen but in 1959 large numbers were recorded. On 8 November, 65 Stock Doves in parties of up to 7 were seen flying out to sea and returning and one party of 18 was seen to come in high. On 15 November, at least 619 Woodpigeons and 87 Stock Doves arrived from the N and NE and turned back N after either reaching the Bill or circling for various distances out over the sea, while to left to the S and 16o to the WSW and smaller numbers of Stock Doves did the same. On 24 November, 14 left SSW at 1240 hours. On 28 November, over 15,000 pigeons (estimated at c. 10,000 Woodpigeons, and c. 5,000 Stock Doves) went over Portland in three great flocks to the SW and SSW out to sea between 0745 and 0815 hours. It is of considerable interest that Woodpigeons were also seen departing from Portland in late summer-14 flew out to sea S and returned on 26 August. Similar observations have been made at other points along the south coast. C. J. Henty (per D. Lack) saw flocks of 390 and 700 pigeons 3 1

27 BIRD STUDY departing S and SSW on 2 November 1957 from Selsey Bill and also saw 300 coasting W and 70 coasting E. At St. Catherine's point, Isle of Wight, on the same day 33o Woodpigeons were recorded arriving from the E and then turned NW or W and these records appear to be of coasting movements. However 52o were seen to depart SSW and another 120 flew out to sea but returned (per D. Lack). Other records for 1959 are summarised in Table XI. Records of immigration on the east coast are also variable in different years. In 1956 a watch was kept from dawn till Iloo hours on the coast one mile south of Seahouses in Northumberland from 1 to 19 November by one of us (M.G.R.). No Woodpigeons were seen at the coast and only small numbers inland. Departures of pigeons from a roost in the early mornings at Kyloe forest showed the same southerly preferences described earlier. A watch was kept on the coast of Norfolk at East Runton, near Cromer from 31 October until 15 November A peak in arrival times of Continental immigrant passerines occurred around o9oo hours (Murton, 1957). Associated with these immigrants three flocks of 20, 12 and 8 Woodpigeons arrived around this time on 7 November from the east. Three Woodpigeons arrived from the east at about o800 hours on 8 November, one from the east on the loth at o800 hours and one from the east at 0830 hours on the 13th. No Woodpigeon activity from inland was noted at the coast such as was observed in Kent. In 1957, three watches were also made at the coast at Sandwich Bay. On 2 November, a watch was made from 130o hours until dusk. The weather was clear and sunny with a strong SW wind and many flocks of migrating Starlings were seen arriving from the E, and when first seen, were well out to sea. At 1322 hours, seven Woodpigeons arrived from the east and when first seen were approximately a mile out to sea. At 1440 hours a single Woodpigeon arrived from the same point. On 2 November a coastal watch was kept from iioo hours until 1530 hours. Again many Starlings and Rooks were arriving and also Skylarks. At 1138 hours, one Woodpigeon arrived from the east and at 1215 hours 3 arrived from the sea flying very high. These small numbers suggest that a small immigration of Woodpigeons may have been taking place but they do not seem likely to have been the cause of the large scale movements on the south coast. In 1959, a different story emerged from coastal observations by one of us (R.K.M.) in Suffolk and from records received from other sources. These records are summarised in Table XI. Thus the autumn of 1959 turned out to be one of large and impressive movements of Woodpigeons. One of the most remarkable of these movements occurred on 20 November when a Pests Officer of the Ministry of Agriculture reported seeing an enormous cloud of Woodpigeons, estimated to be two miles long, arriving from the 32

28 MOVEMENTS OF WOODPIGEONS e - E, a ^ ^ a ^ u a u ^ á 89 ó óó 0 0 bp eog ^, ^ OA^ ^ d ri w w Vw w ^ u 1.. ^ ái às o kl ó 74 `2i às á. u^.1i,.^ p 0 ^ 7 p 7 :750 Op ^ b^ H q a b ^-^+ 0i O y U z ^ ^ ^ó W.-^-, U 04 Q ;" ^ 3 q^clc7cj^ z ó ^ Ó ó ^ o7, >, v o Á.w O ^ ^ N ^ y h s ^-' i ^ M O ^ cd N v a+ ^ O ^ ^ A w zz ^ v 0 0 >^>>>>>>z> U U O yo y O O O yzo yz yo ^j O ^ 0 OOZyZZ ZZZ N z p,.. M M^ vn,nh N fn h.ni 0.1 > > > >z ^ yr-^yyy F-+Frz Nlrz 0,10 ^ N ^ N N N r 33

29 BIRD STUDY sea at Salthouse, Norfolk, during the late afternoon. This observation is confirmed by R. A. Richardson who stated that an immense movement occurred during the latter part of the week ending 21 November. He thought the birds were arriving from the direction of Yorkshire or from further north. The movement resulted in a large increase in local pigeon numbers. Table XI shows that the big recorded movements had four main components. There were arrivals from the E into East Anglia and enormous N-S movements down the entire eastern coast of England. We shall discuss the significance of these later. The other two components consisted of almost equal numbers of departures southwards from the Channel coast and arrivals at it from the south with some coasting movements. Like the movements recorded by Wilkinson (1950), Lack and Ridpath (1955) and our observations in SE Kent, we interpret both components as being typical of departing migrants, some of which were actually seen to depart S, some seen to leave and return and some seen only when they returned, as a result of which they were recorded as `arrivals' from the S. Westerly coasting movements on the south coast appeared to bring some birds to Lands End where departures NNE were recorded. It appears that some crossed St. Georges Channel because there were arrivals from the S into Ireland. Thus J. T. R. Sharock at Cape Atar Island, Co. Cork saw 63 very tired looking Woodpigeons, which arrived from the SSE at 1600 hours on 28 October associated with an arrival into a NNW wind of several other species including Mistle Thrush, Song Thrush, Chaffinch, Redwing, Greenfinch, Skylark and Fieldfare. The observatory closed on 2 November, no more pigeons being seen. At Great Saltee (Co. Wexford) after one Woodpigeon passed on the 3rd, 5oo were seen moving from S to N on the 4th, with a few the next day and 200 on the 6th. There was a further passage on the Loth. If any emigration of Woodpigeons across the Channel to the Continent in autumn were to take place, it would be likely that a return migration would occur in the spring. Very few observations on the spring migration have been made in SE England for any species (Snow, 1953) therefore the lack of records of spring migration of Woodpigeons seemed to be of little significance. In order to see whether any such arrivals took place, a watch was made from dawn at 0500 hours until 1400 hours daily from 19 March to 19 April, 1958 at South Foreland. Twenty-five Woodpigeons were seen to arrive from across the sea flying NW between dawn and 0900 hours, as follows : 3o March, ten; 9 April, one; 10 April, thirteen; and one at 133o hours on 12 April. By themselves these records are no proof of an appreciable migration but they do suggest that a small immigration might have been taking place. The roosts at Waldershare and Tilmanstone contained no Woodpigeons during this period. 34

30 MOVEMENTS OF WOODPIGEONS DISCUSSION Neither ringing nor direct observation gives a complete picture of Woodpigeon movements since the one depends greatly upon chance and the other is very restricted in the area over which it can be used. What follows is a synthesis, which attempts an explanation of the data obtained by both methods. Ringing shows that the Woodpigeon population of Britain is largely sedentary and surprisingly conservative even in its local wanderings which rarely exceed io miles. It is clear that there is no major displacement of the Woodpigeon population of Britain in autumn or winter. Nevertheless, our observations have shown that an increase in the number of birds seen in the early morning can occur in November, and this is also suggested by the results of the 1955 enquiry. All our records have shown that at this time there is a broad tendency for Woodpigeons to fly in southerly and easterly directions in the early mornings in late autumn all over Britain. At first sight these observations seem to contradict the ringing data but this is not necessarily so. It appears from observation of the roost at Carlton that such an apparent increase in late autumn was really an artifact caused by a change in the nature of roost to feeding ground flights and the appearance of juveniles in early November which swelled the ranks of the flocks departing from the roost in the mornings; and the same situation probably applied to some extent to all England. In this connection it is interesting that answers from 54 foresters in NE Scotland to a questionnaire sent out by Robertson (1957), all said that they noticed an increase in Woodpigeon numbers in late November, and 44 mentioned early December as well, and the same effect may be involved. Southerly flights did not in fact mean that actual displacements of the whole population were taking place since it was possible to trace pigeons back in the evenings to the roosts from which they had left the same mornings. These southward flights may represent a latent migratory drive, coming from a much earlier period when British pigeons, like those of Scandinavia, were obliged to migrate south to escape a much more severe winter than the present one in Britain. Amelioration of our climate would have gradually reduced the need to migrate but selection has not yet eliminated this migratory drive altogether and it finds expression in southerly morning flights from the roost in the autumn. However there are grounds for supposing that, although such a migratory drive does no more in most individuals than induce morning flights from their roosts in a southerly direction, among a small proportion the drive is actually fulfilled and the pigeons migrate. This is indicated by ringing which has shown that a small proportion of British Woodpigeons do move appreciable distances in winter, since 25 per cent of winter recoveries were at distances of 26 to 33o miles 35

31 BIRD STUDY from the places in which they were ringed in summer (Table I). Furthermore per cent of the individuals were recovered south of their place of ringing and also 78 per cent were in their first year. It is significant that the percentage of distant recoveries (over five miles from the place of ringing) increases sharply from 8 per cent in September-October, to 56 per cent in November-December. Finally there are four winter recoveries of British Woodpigeons from France, three of them being first year birds. All these facts are in accord with the hypothesis that a small southward movement of British Woodpigeons, consisting of a high proportion of first year birds, takes place in autumn and that in some cases this actually leads to emigration. The beginning of this movement coincides with the sudden increase in the number of first-year pigeons joining the flocks of adults which leave the roost each morning. It is probably this sudden appearance of juvenile pigeons in November coupled with a simultaneous increase in flight activity which is responsible for the widespread belief in the countryside that `foreign' pigeons, being small and dark and lacking white neck marks, arrive in this month. These are really the characteristics of juvenile pigeons and an increase in their abundance is exactly what happens in November. It appears from the coastal watches that in some years, such as 1956, only very small arrivals, or none, occur on the east coast; whereas in other years, such as 1959, recorded arrivals from the east are more substantial. Ticehurst (1932) has described pigeon arrivals from the east on the coast of Suffolk like those we have observed there and in Norfolk, and also saw flights out to sea with returns in autumn (and spring) similar to those on the south coast. He said that such arrivals usually took place from mid-october till late November, and some as late as December and January `in front of or with severe weather conditions'. Witherby et al. (1940) also state that arrivals from the east occur on the east coast once every few years in late autumn. If such arrivals do occur, why have they not been recorded more often, at the bird observatories for. instance? Our evidence suggests a number of reasons why they should not have been seen more frequently. In the first place our observations in Suffolk showed that many of the pigeons seen arriving at the _ coast at a considerable height could easily have been overlooked. The detection of migrants which normally arrive at heights beyond the range of the naked eye is feasible only with the use of radar (Lack, 1959). Secondly, the recorded arrivals, though appreciable, were never on such an enormous scale as in the case of species such as the Starling or Chaffinch which are regularly seen arriving (in spite of the fact that they also frequently arrive high, they are sufficiently numerous for the low arrivals also to be seen regularly). Finally, arrivals of Woodpigeons are not only irregular, occurring once every few years, but also seem to be limited to only 36

32 MOVEMENTS OF WOODPIGEONS a few days, therefore the chances of observing them are relatively slight. The only recovery of a foreign ringed Woodpigeon has already been mentioned. This is in spite of the fact that Woodpigeons are shot in large numbers in Britain and that there are many recoveries of British-ringed Woodpigeons; furthermore this lack of ringing recoveries can hardly be attributed to small chances of recovery since a number of Woodpigeons ringed in Northern Europe, especially Scandinavia, have in fact been recovered (from SW Europe). Apart from the irregular nature of Woodpigeon arrivals from abroad and their relatively small size in most years, this lack of foreign ringing recoveries means that these arrivals from the east are for some reason much less likely to be shot in Britain. It happens that ringing recoveries of British pigeons in November, when the arrivals from the east take place, are the lowest of the year (3.0 per cent) so we assume that they have very little chance of being recovered immediately they reach Britain. We believe that the reason why they are not recovered later is because they have left the country again very quickly and consequently had not remained here long enough to make the chances of a ringed recovery high (but see below). The origin of such arrivals is unknown. The Woodpigeon population of Scandinavia is relatively low and it departs between September and mid-october and arrives at the Pyrénées by the end of October, whereas the arrivals on the British east coast occur later (early November onwards). These British arrivals might well consist of pigeons from more southerly regions of northern Europe as well as of stragglers from Scandinavia. What, then, do these arrivals from the east do when they reach the coast of England in November? The records from the coast show that there is a strong N-S movement down the coast in autumn. The existence of such a movement had also been reported by Colquhoun (195i) who believed that it consisted of Woodpigeons moving from Scotland into southern Britain. The records for November 1959 showed a striking north to south movement of pigeons at Spurn Head, in Norfolk and an even bigger one in north Kent. Although it is probable that this passage contained an element of British pigeons travelling south or south-east, it seems very likely that pigeons arriving from the east turned on arriving at the coast and moved south. Ticehurst (1932) actually saw easterly arrivals turn in this way. We therefore suggest, that any such easterly arrivals either turn and coast southwards, or continue overland finally arriving in Kent and the south coast of England. We have so far suggested two separate but simultaneous southward movements in Britain in November; on the one hand consisting of a general movement south of a small proportion of the British population, including many first year pigeons, and on the other hand 37

33 BIRD STUDY of a passage through SE England of Continental immigrants in some years. We consider that these movements are responsible for the increases of pigeons seen along the south coast in November, in Dorset, Hampshire, Isle of Wight, Sussex and, in particular, for the very large increase in Kent in 1957 and 1959 since it received the stream down the east coast. What happens when these pigeons reach the Channel coast can be deduced from the intensive observations made in SE Kent and from other observations extending along the coast to Cornwall. We now interpret the observations by Wilkinson (1950) and Lack and Ridpath (1955) and those in this paper of pigeons coasting and occasionally making sallies for short distances southwards out to sea and then returning, as being typical of departing migrants. This reluctance to cross the sea is much increased by unfavourable weather such as general haze, fog or headwinds; and this explains the impressive increase in numbers in Kent and along the south coast in November. Some of these southward moving pigeons, which are thus halted, are temporarily drawn into local roosts and cause the build-up in numbers in roosts in Kent. Finally suitable conditions occur for migration across the sea, and they depart. Although slight, we had some evidence for a slight return migration northwards across the Channel to England in late March and early April 1958, but see also Wilkinson (1950). We consider that we have taken the method of direct (visual) observation of movements and roosts as far as it can usefully be taken in this problem. The extent and origin of Continental arrivals and the degree of emigration of British Woodpigeons can best be further investigated by ringing large numbers of Woodpigeons in southern and eastern England in late autumn, preferably using methods of mass capture on flocks, and by the use of radar, work on which is now in progress. SUMMARY (I) Woodpigeon movements in the autumn were investigated by an analysis of B.T.O. ringing recoveries; by an extensive trust aided investigation into the numbers and flight preferences of birds leaving their roosts in the early morning, and by intensive observations over four years at roosts in SE Kent. The records of Woodpigeon immigrants on the east coast of Britain have been examined and supplemented by further observations. (2) Ringing recoveries showed that most British birds are sedentary, 75 per cent staying within 25 miles of where ringed. Among those birds moving longer distances first year birds predominated. A southerly trend was apparent (but not statistically proved) in winter recoveries over five miles. In contrast Continental ringing showed marked migratory movements SW across Europe in the Scandinavian populations. 38

34 MOVEMENTS OF WOODPIGEONS There was little evidence that populations in different parts of Britain showed different migratory tendencies, but for all areas an increase in flights occurred in November. (3) An increase in the number of birds leaving their roosts in November was recorded throughout Britain and their early morning flights were mainly to points between NE and S. This increase was partly an artifact due to the appearance of juveniles at this time. However, in SE Kent genuine increases occurred, which could be explained, in general terms, by arrivals from northerly directions and departures south. The preference for southerly flights was even more apparent at these roosts, and these took the birds to the coast, where all the evidence indicated that they emigrated when weather conditions were favourable. (4) A few Woodpigeons arrive on the east coast of Britain in most autumns but occasionally large numbers have been recorded. These appear to move through SE England causing the increase in roosting totals noted in Kent and eventually to emigrate from the south coast. British birds also show southerly flight preferences at this time which are interpreted as representing a latent migratory drive. This drive is realised in a small proportion of the British population, mainly young birds, and these also move south. In southern England these flights may be confused with the arrival of Continental immigrants and a small number of British birds may emigrate with the true `foreigners'. (5) The origin of immigrants, especially in November, is discussed. They are more likely to come from a more southerly region of North Europe than from Scandinavia. ACKNOWLEDGEMENTS We are greatly indebted to our colleagues, F. J. S. Jones, D. McCowan, D. D. B. Summers, R. J. P. Thearle and E. N. Wright, who gave considerable help in the field during these observations and for their criticism of this paper in manuscript; also to D. Lack and J. S. Ash for valuable advice and information; to M. Westmacott and B. Church for important statistical help and to A. Jenson for preparing the text figures. We are most grateful to the Earl of Guildford and E. Rice for allowing us to make prolonged observations on their properties; and it is a pleasure to thank A. Hogbin for information about Woodpigeons at Tilmanstone. We should also like to thank the Master and Elder Brethren of Trinity House for allowing one of us to be stationed as an observer on board the South Goodwin Light Vessel. Mrs. P. J. Ridpath made translations for us from Italian and Norwegian. Part of this paper is based on the data of many observers to whom we are most grateful, especially for the generally high standard of care with which the observations were recorded. Records and help were received from the following: A. Adams, K. D. Adams, A. F. Airey, L. P. Alder, D. M. Allen, W. T. Allen, W. K. Armitage, G. A. Arnold, C. J. Armour, J. S. Ash, A. G. Ashman, G. Aver, J. A. Bennington, E. W. Bentley, C. Bigley, M. A. Bilby, A. E. Billett, G. E. Black, J. Bound, D. C. Briand, G. A. Bridge, W. H. Brown, P. W. P. Browne, C. E. Burton, Miss S. M. Butlin, G. C. Buxton, D. B. Cabot, J. F. Carter, M. A. R. Cawkell, P 1. Chadwick, R. Chainey, S. H. Chalke, J. Challis, 39

35 BIRD STUDY R. Chestmey, A. Clarke, C. Bennet-Clarke, J. W. Cockburn, D. R. Coda, E. Cohen, B. Coleman, R. J. Colwill, R. Cooke, J. Conchie, J. A. Constable, J. N. Corden, D. M. Cormack, R. S. Cormack, L. A. Cowcill, D. A. Cox, D. M. Crampton, J. Cunningham, J. T. Daye, L. L. Dews, R. F. Dickens, H. Dickenson, T. E. Dodd, D. Drummond, G. K. Dunning, W. H. Edgar, J. A. Elliott, L. Feetham, R. Felton, F. Fincher, B. A. le Neve Foster, J. Gault, H. E. Gengh, W. S. George, A. Gibbs, D. Goodwin, F. W. Green, H. Green, R. J. Green, A. H. Grimsey, D. Guthrie, A. Hall, D. R. Hamlett, E. Hammond, A. J. Hartham, G. Hayes, M. Henderson, C. J. Henty, J. Hewlett, R. A. O. Hickling, D. P. Holmes, W. M. Logan-Home, G. Howlett, P. A. Hoy, G. Hughes, R. Humber, H. G. Hurrell, E. Hutchinson, A. J. Isaacson, D. A. G. Isaksson, V. Johnson, W. Jones, K. M. Kirton, G. Knibbs, M. R. M. Lance, A. C. Leach, H. Lord, J. Lyon, T. Maclennan, V. J. Macrai, H. K. Marshall, J. R. Marshall, W. K. Marshall, H. P. Mayes, W. S. Medlicott, N. C. Moore, N. W. Moore, Mrs. F. Murton, M. T. Myres, H. Naylor, L. Nesbitt, Miss M. G. Newman, Mrs. F. K. Nugent, F. Owen, J. C. Owen, I. Patterson, J. Peake, I. D. Pennie, B. Pitman, G. C. Phillips, R. H. Poulding, P. Prosser, J. A. Pryce, G. H. Reese, P. Reeve, B. A. Richards, J. O. Richards, R. A. Richardson, P. Rickwood, A. F. Ridpath, Mrs. P. J. Ridpath, S. S. H. Rittner, J. A. B. Robertson, R. W. Robson, A. Roebuck, F. Rowe, Mrs. M. Rowe, R. F. Ruttledge, A. Swayer, R. Scott, R. E. Scott, F. C. Skelcher, A. F. Smart, R. K. Smith, J. Stafford, N. F. Stewart, Miss E. Taylor, J. W. A. Thompson, H. V. Thompson, T. A. Toogood, I. F. Trant, E. A. E. Tremlett, J. D. R. Vernon, N. G. Vickers, J. N. Walsh, T. P. WaIsh, T. F. Wardle, A. J. Welling, Miss B. Whiteaker, P. E. Williams, R. F. Wills, F. W. Wood. REFERENCES ALDER, L. P. The Sussex Bird Report for ALDER, L. P. The Sussex Bird Report for ALEXANDER, W. B The Woodpigeon. J.R. Agric. Soc., Ioo: ASH, J. S., RIDLEY, M. W., and RIDLEY, N On the movements and survival of Woodpigeons and Stock Doves. Brit. Birds, BARRIETY, L Le baguage des oiseaux au pays Basque. Bull. Cent. Rech. Sci. Biarritz, I : BARRIETY, L Baguage et reprise d'oiseaux bague en Bull. Cent. Rech. Sci. Biarritz, I : BEESTON, T. J Migration routes of the Woodpigeon in Worcestershire. Brit. Birds, 24: COLQUHOUN, M. K The Woodpigeon in Britain. H.M.S.O., London. FONTAINE, V Góteborgs Naturhistoriska Museums ringmírkningar av flyttfaglar. Góteborgs naturh. Mus., 1 948, 72-97; /52, 45-92; 1953, 9-30; 1 954, 7-26; 1955, 8-29; 1956, GEROUDET, P Les migrations du Printemps Oiseaux, 24 : GODEL, M. and CROUSAZ, G. DE Studien uner den herbstzug auf dem Col de Cou Bretolet. Orn. Beob., HOLGERSEN, H Bird Banding in Norway, Sterna 16, LACK, n Visible migration in SE England, Brit. Birds, 47:1-15. LACK, D Migration across the North Sea studied by Radar. Ibis, lol : LACK, n. and LACK, E Visible migration through the Pyrenees : an autumn reconnaissance. Ibis, 95 : LACK, D. and RIDPATH, M. G Do English Woodpigeons migrate? Brit. Birds, 48: MAYAUD, N L'avifaune des Landes et de la region Pyrénéenc occidentale. Oiseaux et Rev. Fr. Orn., 1940:236-84; 1941,22-38; 1 943, ; 1944, ; 1954, 49'67. 40