The Nesting of the Long-Tailed Tit

Transcription

1 Bird Study ISSN: (Print) (Online) Journal homepage: The Nesting of the Long-Tailed Tit David Lack & Elizabeth Lack To cite this article: David Lack & Elizabeth Lack (1958) The Nesting of the Long-Tailed Tit, Bird Study, 5:1, 1-19, DOI: / To link to this article: Published online: 17 Jun Submit your article to this journal Article views: 502 View related articles Citing articles: 22 View citing articles Full Terms & Conditions of access and use can be found at

2 BIRD STUDY Vol. 5, No. 1, March, 1958 THE NESTING OF THE LONG-TAILED TIT by DAVID AND ELIZABETH LACK (Edward Grey Institute, Oxford) Received 14 June SUMMARY 1. Winter flocks of Long-tailed Tits roosted communally. 2. In each of five years, 5-6 pairs nested on 8o acres, the nests being well spaced, but the species is not territorial. 3. In woodland, many nests were in thorns, usually Bramble or Hawthorn, usually 3-5 feet up, many others were in a bare fork, usually of Oak or Ash, usually feet up, and some were in thornless twigs, such as Clematis or Spruce, usually at an intermediate height. 4. The birds often changed their type of nesting site for repeat nests, which were usually yards away, occasionally yards away from the first nest. 5. Other British observers have found nearly all nests in low thorns, while in Sweden most nests are in Birch or Juniper, and hardly any in thorns. 6. In first attempts the first egg usually follows three weeks after the start of building, but the interval is shorter in repeats, in late repeats being,8-10 days. 7. The first eggs were usually laid at the end of March, but up to three weeks later in a cold spring. Repeat nests follow until the end of April, but rarely later. 8. Clutches were unusually small in the late season of g. The frequency with which three or four parents feed one brood is attributed to the high rate of nest destruction, the restricted breeding season and the absence of territorial behaviour. to. Young were raised from only 16% of the Wytham nests. All in bare forks and thornless twigs were destroyed, nearly all before laying. Of those in low thorns, 12% were successful inside and 47% outside the woodland, and losses were highest during incubation. INTRODUCTION These observations on the Long-tailed Tit (Aegithalos caudatus) were made by E.L. in I and D.L. in , in mixed broad-leaved woodland on the Wytham Estate near Oxford, especially in Marley Wood. The latter is but moderately suitable, since most of the canopy trees of Oak and Ash (a favourite feeding place) were felled in , and much of the wood now consists of overgrown Hazel coppice:, there are also some small marshes and extensive Elder scrub, where Long-tailed Tits rarely occur. Many further nests were found in the Great Wood and two plantations, in which canopy trees are at normal density. In 195o and 1951, most of the individuals in Marley Wood were trapped during the winter and colour-ringed. Personal observations were supplemented by nesting data from the diaries of A. Whitaker, F. C. R. Jourdain and J. H. Owen (partly published in Owen 1945), deposited at the Edward Grey

3 2 D. AND E. LACK Institute, by the B.T.O. Nest Record Cards and by numerous records, by himself and others, kindly sent from Sweden by S. Durango. We are very grateful to Mr. Durango, also to all B.T.O. members who have filled in cards for the species, and to J. Gibb and R. E. Moreau for criticising this paper before publication. PRE-NESTING BEHAVIOUR In at least two winters, two flocks of Long-tailed Tits, each ten or a dozen strong, frequented Marley Wood. Each flock kept to its own part of the wood and they were not seen to meet. When an individual got separated from its flock, it became extremely agitated, calling loudly and flying rapidly about until it again made contact. Tits of the genus Parus often feed for a time with Long-tailed Tits, as is well known in winter, but the attraction continues in March and April when, should a pair of Long-tailed Tits stop building and move into an Oak tree to feed, they are usually joined after a very short time by Parus tits, and sometimes by a Nuthatch (Sitta europaea). Our impression is that the Long-tailed Tits attract the Parus species rather than the reverse. The different species tend to feed on different parts of the trees, and the survival value of the association is probably that the chance of a predator approaching undetected is smaller for a party than a solitary individual. If the trilling hawk-alarm call is given by one of the group, tits of all species respond by immediately diving from the upper parts of the trees to cover below. In different winters we have found five communal roosting places of the Long-tailed Tit in Marley Wood, each involving 9-12 individuals, which in some cases certainly, and in the rest presumably, were the same as the flock which fed together during the day. One roost, used regularly for at least two months, was nine feet above the ground on the bare branch of a Hawthorn. Another, used for six weeks until the birds deserted it through human disturbance, was six feet up in the middle of a dense thicket of Blackthorn ; it was in use again two years later, though not in the intervening winter so far as we know. Another roost, fifteen feet up on a bare Hawthorn branch, was not used after the evening on which it was found, though the birds were not disturbed there by the observer. Two others, found by D. F. Owen following hard after a flock at dusk, were used only once each, probably because of the unavoidable disturbance that his actions entailed. One of them was five feet up in a Blackthorn thicket and the other nearly thirty feet up on a bare Hawthorn branch, the latter seemingly being a lastminute choice of a flock previously disturbed from their normal place by the chasing. When roosting on a branch, the birds usually huddled together in a line, but in a thicket they sometimes appeared to form into a ball, with tails sticking out at diverse angles. Such roosts were found between November and March. One used regularly in March was deserted by each pair successively when their nest was roofed, after which they roosted in their own nest. Elsewhere, E.L. once found a communal roost in June, presumably a family party.

4 1 958 NESTING OF LONG-TAILED TIT 3 At least some colour-ringed individuals nested within the area over which they fed in their winter flock. In March, each pair separates from the flock in order to build, and each also feeds together, but later in the day they may rejoin the flock to feed and for roosting. NUMBERS Wytham Great Wood, with many canopy trees and good undergrowth, is fairly typical of English woodland, and here in pairs of Long-tailed Tits nested on about 65 acres of what is perhaps an optimum habitat. Marley Wood, with few canopy trees and some marshes and scrub, in all about 8o acres (which includes some bordering ground beyond the wood proper), held 6 pairs in 1950, 1951, 1955 and 1957, and 5 pairs in None of these years was preceded by a hard winter. The nearly constant density each year in Marley Wood, coupled with the fact that each pair nested away from its neighbours, might suggest that the Long-tailed Tit is territorial, but it is not. The male does not sing, nor does either of the pair otherwise proclaim itself, or patrol the ground to drive out other Long-tailed Tits. Further, when collecting feathers or feeding, the pair wanders freely over ground which, had the species been territorial, should have belonged to other pairs, and the birds do this without any change from conspicuous to inconspicuous behaviour, such as characterises, for instance, a trespassing Robin (Erithacus rubecula). Both near their nest and on these longer excursions, the pair stays near together, keeping in touch by calls and getting highly agitated should they lose each other; similar behaviour is found in other species which form pairs in flocks, but not in territorial species, which, having a common meeting ground in the territory, do not seem to need contact with - their mates in this way. Moreover the occasional fights between neighbours do not occur at fixed places, such as might suggest territorial boundaries. They happen when one individual comes extremely close to a pair, whereupon one of the pair drives it off, after which it quickly returns to its mate. The impression given is that the bird is defending its mate and certainly not an area. In one instance two building pairs met when each was some 6o yards from their nest, and then all four fought, each of one pair tackling a member of the other. We have also seen strange Long-tailed Tits driven away from the nest, but such chases, like the others, have been discontinued after a short distance. We did not identify the sexes in these encounters. As discussed later, while many repeat nests are built near the destroyed one, three were yards away, and two of these involved the pair crossing over ground used mainly by other pairs, which again suggests the absence of territorial behaviour. So does the unusual frequency with which one and sometimes two extra adults join with a breeding pair, as mentioned later. Indeed, everything suggests that the Long-tailed Tit is not territorial. Because the pairs make rather long excursions for food and feathers, they may well be aware of the position of neighbouring pairs ; and the

5 4 D. AND E. LACK 5(I ) spacing of the nests, and relatively constant density, are presumably due to the birds avoiding building close to each other. Hence an effect, the spacing of breeding pairs, which in other species is ascribed to territorial behaviour, is in the Long-tailed Tit apparently achieved without it. NESTING SITES AT WYTHAM The nests that we found in Wytham differ from the rest summarised in Tables I III in two important respects. First, ours were in broad-leaved woodland, whereas most of the other British ones were in hedgerows and on commons. Secondly, all of ours were found by following the building birds, and hence were random as regards nesting site, whereas most recorded by A. Whitaker and J. H. Owen were explicitly found `by looking in likely places', and so, doubtless, were many of those on B.T.O. Nest Record Cards. Since low nests are far easier to see than high ones, this implies a biassed sample. The Long-tailed Tit nests in three very different situations, in a thorny bush usually from 3 to 6 feet up, in thornless twigs usually from to to 3o feet up, and in the bare fork of a broad-leaved tree usually from 3o to 7o feet up. The records by others suggest that, in England, low thorny bushes are used far more often than anything else, but these samples were biassed, as already mentioned. Our own records for Wytham Great Wood and adjoining plantations, which approximate to natural woodland, show that nests under to feet and over 20 feet up were almost equally common. All three types of nesting site occurred in the Great Wood, but nests in thornless twigs were few, and unusually high, nearly all being in Spruce (Picea abies), which is an introduced tree ; and in that part most like natural woodland, the height of nesting was effectively bimodal, nearly all nests being either more than 3o feet up in the fork of an Oak (or Ash) or less than 6 feet up in Brambles. Marley Wood did not show this bimodal effect because nests in canopy trees were scarcer, while most of those in thornless twigs were at an intermediate height. The most popular low site, including 4o% of all nests found in Wytham, was in Bramble (Rubus spp.). The nest was 2 to feet, usually 3 to 5 feet, off the ground, often about a foot below the top of a thick clump and from 2 to 6 feet in from the side, and it was closely woven with the prickly stems. Some, though not most, nests had a thorny stem close in front of the entrance ; this might have been accidental, but makes it harder for a predator to enter. One nest in Bramble was placed against a thick Ash trunk, and two were screened in front, one by thick Honeysuckle (Lonicera periclymenum) and the other by dense low hanging branches of a Larch (Larix decidua). Hawthorns (Crataegus monogyna) are almost as common as Brambles in Wytham, but usually grow less thickly, and Longtailed Tits used them less often, in all 15 % of the nests being in Hawthorn. The nest was placed either against the main stem or amid the twigs, and was more accessible than in Brambles, except in the few places where the bushes have been clipped in a hedge or

6 1958 NESTING OF LONG-TAILED TIT 5 (in the Park) eaten back by cattle, where the growth is denser and the nests harder to reach, such sites being very popular. Blackthorn (Prunus spinosa), though not uncommon, grows chiefly in loose thickets and was hardly ever used. One nest in a Blackthorn thicket, and another in a Blackthorn hedge, were actually placed in a Bramble and Hawthorn respectively, have been classified under these latter species. Wild Rose (Rosa canina), though common, is usually thin and straggly, and was used only once. The only clump of Gorse (Ulex europaea) was cut down late in 1955, but, though small, was used twice in 3 years. At Wytham, thornless twigs were the least popular of the three main sites. Five such nests were in Spruce, placed among the needles on a side branch, often partly below the branch, at from 25 to nearly 70 feet above the ground. Four others were in, or just behind, dense hanging strands of Traveller's Joy (Clematis vitalba), two being amid the stems of the Clematis itself, at 15 and 25 feet above the ground, and two others in the fork of a tree covered by the Clematis, one in a thin Hawthorn at 25 feet and one in a low Elder (Sambucus niger) at 32 feet. Another nest among thornless twigs was in Privet (Ligustrum vulgare) at so feet, and another, most unusually, in the terminal twigs of an Ash (Fraxinus excelsior) at 50 feet. Of the nests in bare forks, most were in Oak (Quercus robur). 15% of all nests found in Wytham were in Oaks. Ash was also used not uncommonly, and one nest was in a Wych Elm (Ulmus glabra). These nests were from 30 to 70 feet up, usually in the fork between the trunk or a large vertical branch and a side branch, but twice on a ledge from which a side branch had broken off, the nest being supported on only one side. The other common tree in the canopy, Sycamore (Acer pseudoplatanus), was not used. One nest was so feet up in the fork of an Elder, a common species, but otherwise used only for the very low nest, already mentioned, screened by Clematis. One further nest was in the bare open fork of a dead Larch sapling, only 3 feet up, quite unscreened, and highly conspicuous. TABLE I HEIGHT OF LONG-TAILED TITS' NESTS ABOVE GROUND i Ht. in ft. Wytham Estate England Sweden Great Wood Marley (B. T.O. cards) Ht. in metres (Durango) Total NOTE: Wytham nests were a random sample

7 6 D. AND E. LACK S(I) Type of Site TABLE II NESTING SITES OF LONG-TAILED TIT Great Marley Wood Wood --(personal) Boars Hill Derby Salop England Sweden Yorks Mon. (Whitaker) (Owen) (B. T.O.) (Durango) A. Low THORN Bramble (Rubus) Hawthorn (Crataegus) Blackthorn (Prunus) Wild Rose (Rosa) Gorse (Ulex) Holly (Ilex) Garden exotics B. THORNLESS TWIGS Broom (Cytisus) 3 Honeysuckle (Lonicera) 3 Clematis (Clematis) 4 Privet (Ligustrum) Ivy (Hedera) Box (Buxus) Laurel (Laurus).. 1 Elder (Sambucus) 1 Ash (Fraxinus).. 1 Scots Pine (Pinus) 1 Spruce (Picea) Cypress (Cupressus) 1 Yew (Taxus) Larch (Larix). 1 Juniper (Juniperus) 29 Dead brushwood 1 C. BARE FORK Oak (Quercus) Ash (Fraxinus) Elm (Ubnus) Alder (Alnus) Willow (Salix) Birch (Betula) Beech (Fagus)... 4 Aspen (Populus).. 4 Rowan (Sorbus) 1 Elder (Sambucus) Larch (Larix)... 1 Fruit Tree (Pyrus) 2 NOTE: The Swedish records were given only for the species of tree, not as to whether the nest was in a fork or among twigs. Two further Swedish records were in nesting boxes. A Long-tailed Tit's nest is hard to see in the high fork of an Oak, as its covering of lichens matches the lichens on the trunk. But in Brambles, which are leafless when the birds build there, the same lichen covering makes the nest conspicuous. It does not necessarily follow that a bare fork was the ancestral site of the species, for concealment is specially valuable on an exposed Oak tree, whereas in Brambles the thorns provide an alternative protection ; and one could hardly expect the evolution of two different types of nest, adapted to these different sites.

8 i95 8 NESTING OF LONG -TAILED TIT 7 TABLE III COMPARISON OF NESTING SITES IN DIFFERENT AREAS Place Total found Percentage found in A loer thorn thornless twigs Mare fork Great Wood Marley Derby-Yorks Salop Mon England (B.T.O.) Sweden OTHER BRITISH RECORDS Open broad-leaved woodland was presumably the natural habitat of the Long-tailed Tit in England during the Atlantic period. The species is also regular at all seasons in pure coniferous plantations, as on the Breckland in Norfolk, where J. A. Gibb has found it nesting commonly 20 to 3o feet up in Scots Pine (Pinus sylvestris) and Corsican Pine (P. nigra), and also a few feet up in Broom (Cvtisus scoparius). But at the present day many, and perhaps most, English Long-tailed Tits nest in hedgerows, also in commons and wild gardens, and here, though the samples in Table II found by A. Whitaker, J. H. Owen and B.T.O. members were probably biassed, low thorns are evidently used far more commonly than either thornless twigs or high forks. The moss popular bush was Hawthorn in the B.T.O. records, Blackthorn in J. H. Owen's sample and Gorse in A. Whitaker's records and our few personal ones for Boar's Hill, near Oxford. Hawthorn, Blackthorn and Wild Rose grow much more thickly, and so provide more suitable sites, in hedges than woods. Most of the B.T.O. nests were 3 to 6 feet above the ground, several only 2 feet up. Details of other plants used are given in Table II. The B.T.O. sample as yet includes only 3 nests from Scotland, of which 2 were in bare forks in Birch trees (Betula sp.). Similarly B. Campbell informs me that of 8 nests that he has seen in Argyllshire, i were in high forks, 4 in Birch, 3 in Oak, while the eighth was low in a Beech hedge. NESTING SITES IN SWEDEN The Swedish records from various observers compiled by S. Durango may not have been a completely random sample, but they suffice to show big differences from England. Only 1% were in low thorns, as compared with 51% in Wytham Great Wood and 86% in the B.T.O. sample. Further, the height records were distributed fairly evenly between I and 15 metres, with 6 to to metres most frequent, so that there was neither the bimodal distribution found in Wytham Great Wood, nor the predominance of low sites found in English lanes. The Swedish nests were either in a bare fork or among thornless twigs, but the two were not distinguished (hence the summary for Sweden in Table II refers solely to the

9 8 D. AND E. LACK 5(I) kind of plant). Birch was much the commonest tree and Juniper (Juniperus communis) the commonest shrub. Some nests were alongside nests of the Wood Ant (Formica rufa) and others amid colonies of the Fieldfare (Turdus pilaris), which presumably give protection from predators (Durango 1 949). The marked differences in nesting between Swedish A.c. caudatus and English A.c. rosaceus may well be due wholly to differences in the opportunities provided by Swedish and English habitats respectively, and there is no need on present evidence to attribute them to hereditary differences. In support of this view, Scottish A.c. rosaceus often use forks in Birch like A.c. caudatus in Sweden, and English A.c. rosaceus nest in the introduced Spruce, although it provides a very different nesting site from those in natural English woodland. Summarising, the Long-tailed Tit has more varied nesting sites than any other British passerine species, and it would be hard to match it anywhere in the world. The nest may be from 2 to 70 feet above the ground, and deep in a thorny bush, or among thornless twigs, usually evergreen but sometimes bare, or in a bare open fork, large or small. Finally, protection may be given by inaccessibility on a terminal branch, or by a lichen covering which matches the lichens on a trunk, or by a screen of creepers or dense conifer needles, or by numerous thorns, or by proximity to the nest of a fiercer bird or insect. As an abnormality the species may even breed in a nesting box (per S. Durango) or in a hole in the ground (Pfeifer 1938, Thielemann 1956). SITE OF REPEAT NESTS BY SAME PAIR Does the same pair keep to the same type of nesting site? In Marley Wood in 195o and 1951 the owners of repeat nests were TABLE IV REPEAT NESTS BY SAME PAIR AT WYTHAM 1st Nest 1st Repeat Later Repeats 1950 (i) Bramble 3 ft. Wild Rose 7 ft. Hawthorn 3 ft. {Oak fork 30 ft. (ii) Spruce twigs 16 ft. Oak fork 20 ft (iii) Elm fork 40 ft. Bramble 3 ft. (iv) Oak fork 20 ft. Oak fork 35 ft. Elder fork 10 ft. Oak fork 35 ft. (v) Hawthorn 4 ft. Privet twigs 10 ft (vi) Clematis 24 ft. Bramble 4 ft. (vii) Clematis 15 ft. Bramble 4 ft (viii) Bramble 5 ft. Bramble 5 ft. (ix) Bramble 7 ft. Bramble 21 ft. Hawthorn 4 ft. (x) Clematis 25 ft. Hawthorn 4 ft. (xi) Oak fork 50 ft. Bramble 3 ft. (xii) Ash fork 30 ft. Bramble 3 ft. (xiii) Ash fork 30 ft. Bramble 3 ft. NOTES: (i) All these were in Marley Wood except for the last three, which were in Great Wood or adjoining plantations. (ii) A. Whitaker (unpublished diary) recorded a pair which first nested 5 ft. up in a Wild Rose, then in the fork of an Oak at 25 ft.

10 1958 NESTING OF LONG-TAILED TIT 9 identified by their coloured rings, while various other repeat nests could be confidently ascribed to particular pairs through their proximity to the destroyed nest, taken in conjunction with the dates involved and the circumstances of all adjoining pairs. The results in Table IV show that a pair often changed its type of nesting site in successive attempts, though some pairs, notably (iv) in 1951, did not. Pair (i) was unusual in that, after their first nest in Brambles was destroyed, they started building simultaneously in a Wild Rose and the high fork of an Oak, some ten yards apart, but the Oak was quickly abandoned for the Rose ; after the latter nest was destroyed, they re-nested in a Hawthorn. The repeat nest was often 5o to 8o yards from the destroyed one, twice only 40 yards away, but not closer. At the other extreme, in 1951 one colour-ringed pair repeated nearly 700 yards away, crossing woodland occupied by another pair. To judge by the fitting of dates and the circumstances of other pairs, another pair moved 600 yards for their repeat in 1955, while in 1957 pair (viii) put their first repeat 40 yards away, but their second 55o yards away, past ground occupied by 2 other pairs. Most shifts of yards would take the birds outside our study area, and since we failed to trace quite a number of other pairs after their nest had been destroyed, such long shifts are possibly not infrequent. TIME SPENT IN BUILDING In 1955, 3 first attempts were found just after building had begun, and they received their first eggs respectively 17, 20 and 25 days later. Two other pairs had their first nests destroyed after they were domed, and the interval between the start of the repeat and the laying of the first egg in it was in each case only days. In 1957, with a sudden warm spell, nearly every pair started building for the first time on II March. In 5 first nests found just after I I March, and almost certainly started then, the first egg was laid days later, with a mean of 21 days. In 2 nests, one certainly and the other presumably a repeat, started about 18 March, the first egg was laid 18 days later, but in 3 later nests the interval was much shorter. At first repeats started on 30 March and about 6 April the first egg followed respectively io and 8 days later, while another pair lost their second attempt on 8 April and the first egg in the next repeat came Io days afterwards. That building is so much quicker in late than first nests is, we suggest, because the first nests are so timed that the young are present when food is most plentiful, hence the best chance of feeding a late brood successfully is to hatch them as soon as possible. Brown (1924) thought building took 18 days, but did not distinguish first from repeat nests. Owen (1945) found 21 days the average time for first nests, while one repeat took z 1 days. The longest published record of building is for a nest started on 24 January, in which building continued for at least six weeks, the young were still present on 8 May but probably left next day (Tracy 1932). If the clutch was of Io eggs, the incubation period of 14 days and the nestling period of 18 days, the first egg was probably laid on

11 IU D, AND E. LACK5(1) 29 March, 9 weeks after building started. In the Great Tit, like the Long-tailed "Tit, building takes a much smaller number of days in late than first nests (J. A. Gibb, in Litt.). UNUSUAL CONSTRUCTIONS In one second repeat in 195o, the start of which we missed, the birds hurried so much that the nest was still undomed on I I May, although by then it held its fifth egg; it was robbed soon afterwards. In another nest, of the normal pattern when built and until the young were several days old, the roof later collapsed and the young thereafter sat in an open cup ; they were reared successfully. Another nest had an extra hole in the top, through which the brooding bird's tail projected (cf. Robertson & Porter 1952). Another became tipped completely sideways, so that the long axis was horizontal, but the young were successfully raised. A few of those in high forks were blown loose during gales. Owen (1945) recorded other cases of structural damage to nests. DATE OF LAYING IN FIRST NESTS During laying, one egg is normally laid each day, as we repeatedly checked. This means that the date of the first egg can be established by feeling into each nest once during laying, preferably before there are more than 4 eggs. Our results for first nests (i.e. omitting all repeats) are set out in Table V. The 6 years have been of 3 types, (i) 1949 and 1957, when laying started about the end of \larch, (ii) 1910 and 1956, when it started about 9 April, and (iii) 1951 and 1955, when it started about 21 April. Year No. TABLE V DATE OF FIRST EGG IN FIRST ATTEMPTS (WYTHAM) Long-tailed Tit Mean Date Limits Great Tit Mean Date Difference in means March 28 March 1 April 23 April 24 days April 9-11 April 29 April April April 10 May April April 4 May April 5-11 April 4 May April 30 March-3 April 17 April 16 The B.T.O. Nest Record Cards provide information for , but are useless for calculating an average date of laying, since repeat nests were not distinguished from first attempts. Using the cards for England south of the Wash and ignoring a few isolated nests started long before the rest in the year in question, the first egg s were recorded on 3o-31 March in 1948, 1949, 195o and 1952, on 2-4 April in 1953 and 1954, on April in 1951 and 1956, and on 17 April in In southern England, therefore, laying may start at the end of March or in the first days of April, but is occasionally delayed for from 1 to 3 weeks, as in 1951 and 1955, which were unseasonably cold in late March and April. Laying started exceptionally early in 2 nests found by B.T.O. observers in

12 1958 NESTING OF LONG-TAILED TIT II 1952, in one of them on 23 February (C. G. Bennett) while in the other the young flew on II April (E. F. Crosby), meaning that the first egg was laid around 3 March. Table V shows that in Marley Wood the first clutches of the Great Tit have each year been started between 2 and 4 weeks later than those of the Long-tailed Tit, the interval being shortest in 19J5, when the Long-tailed Tits were unusually delayed by cold weather, and longest in 1956, when unusually cold weather supervened after the Long-tailed Tits had started but before the Great Tits did so. Presumably Long-tailed Tits depend for feeding their young either on different insects to the Great Tit, or on the same insects (defoliating caterpillars) at an earlier stage, when they are smaller, but we have not studied this point. DATE OF LAYING IN REPEAT NESTS Laying occurs in repeat nests through April, but we have found only 3 nests later than this, the roofless one already mentioned in which laying began on 7 May, 1950, a nest being built which had reached the cup stage on 5 May, 1951, and another being lined with feathers on 3 June, 1951, both of which were destroyed before laying. In 195o and 1951, special watch was kept in Marley Wood during May, and although only 1 out of 6 pairs in 1950 and 3 out of 6 pairs in 1951 raised young, we found no trace of further repeats by the other pairs in May and think that they probably did not build again. (In 1955 and 1956 search was not maintained during May, while in 1957 most pairs in Marley Wood had young in May.) Lqngtailed Tits without young become extremely inconspicuous in the woods in May, and we do not know what happens to them. The B.T.O. Nest Record Cards likewise suggest a marked drop in new clutches after the end of April. For England south of the Wash, in the years inclusive, there were 112 records of nests in which the eggs were counted during laying, enabling the first egg to be precisely dated. In 49 nests laying began before 16 April, in 53 nests between 16 and 3o April, and in only jo nests between 1 and 12 May, with none later. Moreover 7 of the io May records were in the unusually late and cold spring of 1951, when almost all first nests were started in the second half of April, so that these 7 were not unusually late relative to the start of breeding in that year. Many other B.T.O. cards gave enough information to determine the month in which laying started. Adding these to the others, laying started before the end of April in 172 nests and during May in 14 nests (and 3 more were probably started in late April and 4 more in early May). Hence laying started in May in only some 18 out of 193 nests, or 9%, while if the late season of i951 is excluded, the figure is reduced to 9 out of 147, or 6%. Likewise the dated clutches for England south of the Wash in the diaries of F. C. R. Jourdain include only I out of 29 started in May. The sample of nests found by B.T.O. members is probably somewhat biassed as to date, since Long-tailed Tits' nests are much easier to find by inspection before than after the leaves come out, and further, B.T.O. members appear to search more actively in April than May. Thus there is a marked drop in the number of Nest

13 I2 D. AFD E. LACK Record Cards for the Blackbird (Turdus merula) and Song Thrush (T. ericetorum) in May as compared with April (Myres 1955), and this does not reflect what happens in nature. But the drop in these thrushes is much smaller than in the Long-tailed Tit. Further, there is no drop in the number of Long-tailed Tit records in late as compared with early April. We conclude that in southern England only a few Long-tailed Tits re-nest after April, though the proportion doing so may be rather higher than indicated by the Nest Record Cards. It may be added that the speed with which repeat nests are built in April suggests that breeding as late as May is disadvantageous. A. Whitaker's records for the Derby Yorks. border, and the rather few B.T.O. Nest Record Cards for England north of the Wash, show that here, as was to be expected, Long-tailed Tits tend to start and continue breeding rather later than further south, and nests up to the middle of May are not uncommon, with a very few towards the end of the month. Jourdain (1938), apparently based on Brown (1924), stated that there are occasional very late nests. As already mentioned, one nest in Marley Wood was nearing completion on 3 June, while A. Whitaker recorded one with 5 fresh eggs in Shropshire on 16 June (but, as he did not mention whether he saw the parent birds, this was conceivably a deserted nest begun earlier). We have no further records of June nests (one B.T.O. record was clearly erroneous). CLUTCH Feeling in with one finger, we found it impossible to be sure of the number of eggs present when it exceeded 7, which it did in most years. But of 5 clutches in the late season of 1955, four consisted of seven eggs and one of six. The Great Tit likewise laid smaller clutches in a late year (Lack 1955). F. C. R. Jourdain's diaries record only 7 clutches of the Longtailed Tits in southern England (1/9, 2/1o, 4/11). A. Whitaker's records, set out in Table VI, show that in Derby Yorks. the usual clutch is 9-11, with an average near 10. But the average for the B.T.O. Nest Records, also set out in Table VI, is probably too low, since some observers felt in with one finger, which is likely to result in an underestimate. Brown (1924) considered the normal clutch in Cumberland to be 8-12, occasionally 5 or 6, but gave no dated records. The figures in Table VI for Sweden show that the average here is rather higher than in northern England. In England, it is generally believed that clutches of more than 1 2 eggs are due to two hens. The dates of laying in a B.T.O. nest with 14 eggs showed that on at least two consecutive days, 2 eggs were laid each day, which almost certainly means that two hens were involved. A more definite instance of two hens laying in the same nest was observed by A. Whitaker. Laying started on zo April ; on 3o April 4 adults were put out of the nest at dusk, which then contained 6 eggs, which were taken ; on 13 May it contained 9 more eggs, 4 of which were larger than the others, so presumably two hens had laid there. In some of the instances discussed later of more than two adults at a nest, the clutch or brood exceeded 12.

14 1958 NESTING OF LONG-TAILED TIT 13 TABLE VI CLUTCH SIZE Clutch Wytham 1955 S. England (B.T.O.) Number of records Derby-Yorks (Whitaker) Sweden (Durango) Mean VISITS TO THE NEST In 195o and 1951, E.L. watched for periods of up to 3 hours at various stages of the breeding cycle. During building, the pair normally collected material near each other and flew back to the nest together, each building its own piece in. In one colour-ringed pair, the leading bird was usually the male but occasionally the female. Sometimes, one of the pair comes alone with material and, at another nest, colour-ringing showed that this was always the female. Building occurs more rapidly in warm than cold weather, in the morning than the afternoon, in the early stages of the nest than when feathers are being brought to line it, and in repeat than first nests. Counting separately each visit by each of the pair, the fastest rate of building was at a late repeat nest on 5 May, when 74 visits with material were made in 93 minutes, or 48 per hour. At a first nest, the birds visited 33 times per hour over 22 hours on 24 March and 34 times per hour over 3 hours on 25 March, while in other watches at this and other nests, the rate varied between 8 and 25 visits per hour. These figures refer to morning watches at nests before they were lined. At the repeat nest just mentioned, one week later, the birds paid only 16 visits with feathers in 15o minutes, a rate of 6 per hour, and at a first nest at the same stage the average rate was only I2 visits per hour. The birds travel much further to seek for feathers than for other material, and when one of them finds a feather, it often calls excitedly, which brings the mate up to Look, and both then go off to the nest. Because the birds then travel further and build less frequently, the nests are much harder to find when being lined than earlier on. The pair continues to bring feathers during the laying period. Observations of two colour-ringed pairs during incubation showed that, as in Parus tits, only the female incubated, while the male regularly visited the nest to feed the female, and sometimes the female left to get food for herself. In 22 hours at one nest, the male brought food to the female on the nest 6 times, and the female left

15 14 D. AND E. LACK 5(I) the nest herself 5 times, for periods lasting 5 to 15 minutes (and totalling 47 minutes). In the course of one hour at another nest, the male fed the female 9 times, 8 times when she was on the nest and once when she was off, and the female left the nest on two occasions. A few morning watches at nests with young gave feeding frequencies of 25 to 37 visits per hour. The parents brought both very small caterpillars and adult insects, most, if not all, of which are found in the canopy. When fairly old young defaecated, they sometimes clung upside down from the roof of the nest with the long axis of the body parallel with the ground and the cloaca pointing towards the nesting hole, the parent then taking the faeces from them while perched just outside the hole. MORE THAN TWO ADULTS AT ONE NEST At one nest in Marley in 195o, both the colour-ringed parents and an unringed adult were seen feeding the young on ix May. The strange adult was not seen there during incubation or when the young were newly hatched on 3o April. In 195o most adults in the wood were ringed, but the male of the pair adjacent to this nest was not. This adjacent nest was destroyed at the end of April, and it may well have been the male from it which, after its own breeding was frustrated, attached itself to the successful pair. The Long-tailed Tit is notorious among British birds in the frequency with which more than two adults are seen feeding young in the nest. The magazine British Birds alone has included 12 references involving 15 instances (in vol. 1 :32, 62 ; 4 : 78, 209 ; 22 : 38 ; 29 : 80 ; 35 : 59 ; 36 : 55 ; 38 : 272 ; 44 : 388 ; 45 : 257 ; 48 : 92). Since some of these records came in groups, one stimulating readers to send others, it seems probable that similar observations passed unrecorded in intervening years. In 6 and probably 2 more of these instances, the clutch or brood was of normal size (8-1 r), but one nest contained 16 eggs and another about 20 young, presumably the product of 2 hens (British Birds, 29 : 8o ; 38 : 272 ; see also note 36 : 55). In I r instances 3 adults, and in 4 instances 4 adults, were involved. Most were recorded feeding young, but at least 3 instances occurred with eggs. A. Whitaker's record of 4 adults in one nest during laying was mentioned earlier. In a German instance with a clutch of normal size and originally 2 parents, certainly 4 adults and apparently 6 were seen at the nest later (Thielmann 1956). The unusual prevalence of this habit in the Long-tailed Tit is, we suggest, correlated with three factors. First, the extremely high nesting losses mean that many pairs are frustrated in their own breeding. But this by itself is not enough, since nesting losses are likewise extremely heavy in such species as the Blackbird and Song Thrush, which do not show this behaviour. Blackbird and Song Thrush, however, continue breeding successfully through April, May and June, whereas Long-tailed Tits do not usually lay again after the end of April ; hence we think that a second important factor is the restricted breeding season. Thirdly, the absence of territorial behaviour may be important. The first two of these factors could especially account for extra adults attaching themselves to

16 1958 NESTING OF LONG-TAILED TIT I5 parents feeding young. The subject should be studied further in colour-ringed birds, especially in regard to the laying of two females in the same nest. TABLE VII NUMBER OF SUCCESSFUL AND DESTROYED NESTS AT WYTHAM First attempts repeats per cent Situation succeeded destroyed succeeded destroyed successful Bare fork Thornless twigs Thorns in wood Thorns at edge Year succeeded destroyed succeeded destroyed First attempts Repeats TOTAL NOTE: All these nests were found during building and followed through to their fate. NESTING LOSSES Our figures for nesting losses on the Wytham Estate are set out in Table VII, for all those nests which we found during building and followed through to failure or the successful raising of ypung. We took care not to disturb the nests and think it most unlikely that our visits increased their liability to predation, while no losses were due to small boys or other human causes. In all, 56 out of 67 nests were destroyed before the young left, a loss of 84%, which was higher than that recorded for any other species of bird (cf. Lack 1954). However, D. W. Snow (1958) has found that 86% of the Blackbird nests in Wytham Woods were destroyed by predation. In the Long-tailed Tit, there were no clear differences between different years as such, or between first and repeat nests as such, but marked differences between different nesting sites. None of the 15 nests in bare forks, and none of the 9 in thornless twigs, were successful, whereas of those in low thorns II out of 43, or 26%, were successful. Actually, of the nests in thorns in the woodland proper, only 3 out of 26, or 12% were successful, as compared with 8 out of 17, or 47%, in thorns at the wood-edge (the latter including 8 in isolated Hawthorns and Brambles in the Park, 5 in an area of Hawthorn and Bramble scrub with hardly any trees, 3 in irregularly clipped hedges between wood and field, and one in an isolated clump of Gorse). If these latter nests are excluded, only 3 out of the 5o nests in the woodland proper, or 6%, were successful, but their exclusion is not justifiable since the birds in question fed in the woods, though nesting at its edge. Table VIII shows the stage at which each nest was destroyed. Omitting the few nests visited too infrequently for precise dating, all I I nests in bare forks were destroyed before laying (if we assume

17 16 D. AND E. LACK 5(1) TABLE VIII STAGE AT WHICH NEST WAS DESTROYED number destroyed percent destroyed bare thornless low thorns of those at risk forks and low Stage forks twigs in wood at edge twigs thorns Earlier in building Shortly before laying During laying (100) 9 During incubation With young Not recorded SUCCESSFUL NOTE: The percentage destroyed of those at risk is calculated with reference to the total available, including any successful nests but omitting those in which the stage at which destruction occurred was not recorded. Thus for nests in thorns, 42 were available initially, of which 8, or 19%, were destroyed before laying; this leaves 34 of which 3, or 9%, were destroyed during laying; this in turn leaves 31, of which 16, or 52%, were destroyed during incubation, and so on. that laying would have occurred in them when it did in comparable nests in low thorns). Similarly, six out of seven nests in thornless twigs were destroyed before laying, and the other during laying-..on the other hand, of 31 nests in low thorns, only 8 were destroyed before laying, three during laying, as many as 16 during incubation, and another four with young. The best measure of survival is to take into account the successful nests, and to calculate the percentage destroyed of those at risk in the period in question, the method being explained in the note to Table VIII. Of the nests in bare forks and thornless twigs combined, 94% (17 out of 18) were destroyed before laying, as compared with only 19% (8 out of 42) of those in low thorns. For the latter, incubation is evidently the most dangerous period, 52 %, of the nests then at risk being destroyed (incubation lasts about 14 days), as compared with g%o destroyed during laying (which lasts about 9 days) and 27% destroyed during the nestling period (which lasts up to 18 days). The loss of all nests in high forks could hardly be typical for England as a whole, or the habit of nesting there would presumably have been eliminated by natural selection. The explanation is perhaps that, through recent forestry operations in Wytham, nearly all secondary growth and many of the smaller trunks and lower branches have been removed, so that the high forks on the remaining trees may be much more exposed than they would be in natural woodland. Probably two factors account for the much greater losses in low thorns inside than outside the woodland. First, the bushes are thicker outside the woodland, Brambles growing in taller and denser clumps away from trees, and Hawthorn twigs being thicker where clipped by man (in hedges) or cattle (in the Park; hence such edge nests are harder to see and less accessible, at least to man, than those in the woodland. But the difference in the rate of

18 1958 NESTING OF LONG-TAILED TIT 17 predation is so marked as to suggest, in addition, that the chief predator is a species which usually keeps to where there are trees overhead. Much the most common type of destruction in all types of nesting site was for the top of the nest to be removed, with the feathers of the lining scattered about. This indicates the work of a fairly large animal. Further, since nests in both high forks and low thorns were destroyed in this way, the predator was almost certainly a bird, since few mammals could climb the bare trunks and no largish mammal could easily penetrate to the nests in dense thorns. We think it almost certain that the Jay (Garrulus glandarius) was responsible, since it is numerous in Wytham, much prefers woodland to the Park or thorn scrub, could reach all the nests, and could damage them in the way described. Of other possible enemies, the Magpie (Pica pica) is scarce and prefers the Park and thorn scrub (where few nests were destroyed). Grey Squirrels (Sciurus carolinensis) are also scarce, because shot, and could hardly have reached the nests in thorns, but may have taken at least one nest in a high fork, of which the entrance was much enlarged and the body disturbed. Three nests in thorns, otherwise intact, were entered by a small hole made in the roof or back, and in two others the eggs disappeared without visible disturbance, and these five losses we attribute to mice or voles. (A. Whitaker found a vole in one of his nests.) Birds'-nesting boys are effectively excluded from Wytham. Parent Long-tailed Tits are extremely conspicuous during the main period of building, but pay far fewer, though still, rather noisy, visits when lining the nest with feathers ; they become extremely inconspicuous during laying, remain inconspicuous during incubation, and visit frequently but rather quietly when feeding young. Hence if the chief natural predator finds nests by watching the parent birds, one would expect most nests to be destroyed during building before lining, and a fairly large number during the feeding of the young, but few during incubation and extremely few during laying. The times at which nests were and were not destroyed are contrary to this expectation. Thus a large number were destroyed after building was over except for the addition of a few more feathers, a not insignificant number were taken during laying, proportionately more were destroyed during incubation than at any other time, and fewer with young. Further, if the predator followed the building birds, one might have expected nests in high forks and low thorns to be equally vulnerable, whereas far more in high forks than low thorns were destroyed during building. We therefore conclude that the chief predator did not usually find the nests by watching the parent birds, but by searching the vegetation. It fits with this view that comparatively few nests were destroyed with young, since by then the Ieaves are out and the nests are much more hidden. The high proportion of nests opened up before laying also indicates that the predator does not know, when it finds a nest, whether or not it contains anything edible. Of 7o Long-tailed Tits' nests in hedgerows and lanes recorded by Owen ( 1 945), 45 (64%) were successful, a far higher proportion

19 I$ D. AND E. LACK 5(I) than in Wytham, though it is not quite clear whether all of Owen's nests were followed up from the time of building. Unfortunately the B.T.O. Nest Record Cards cannot be used to analyse nesting losses, owing to the infrequency and unpredictability of observers' visits. To instance only one difficulty, of nests visited only twice after a fairly short interval, those found destroyed on the second visit can be recorded as such, but those still intact remain indeterminate. A very rough count of the cards for the years included some 45 successes and 71 failures, which, however it be interpreted, suggests a much higher rate of success than at Wytham. Nearly all of these nests, like Owen's, were in low thorns away from woods, so that despite one additional enemy, the birds'-nesting boy, hedgerow nests are evidently much more successful than woodland ones. This supports the view that Jays are the chief enemy. The rate of loss in Wytham was so high that it may be wondered whether the Long-tailed Tit population can successfully replace itself there. But though a few eggs fail to hatch and a few young die in the nest, each successful brood is large. Further, that only 16% of the nests are successful does not mean that only 16% of the pairs raise a brood, since those that fail, and especially those that fail early in the season, try again, regularly once, quite often twice and sometimes 3 times. Probably, therefore, the woodland population can maintain itself, provided that those at the woodedge are included. Finally, if the later nests are much better concealed, because the leaves have come out, it may be asked why most Long-tailed Tits cease to make further nests at about this time. The answer, presumably, is that while such nests would probably be safer from predators, the parents would not be able to find enough food for their young. REFERENCES BROWN, R. H The nest-building and other breeding habits of the Long-tailed Tit. Brit. Birds, 17: DURANGO, s The nesting associations of birds with social insects and with birds of different species. Ibis, 91 : 14o-3 JOURDAIN, F. C. R In The Handbook of British Birds, ed. H. F. Witherby et al., 1 :272. London. LACK, D The Natural Regulation of Animal Numbers. Oxford. Pp LACK, D British Tits (Parus spp.) in nesting boxes. Ardea, 43: MANSE, V Four Long-tailed Tits feeding young in nest. Brit. Birds, 36 : 55 MvRES, M. T The breeding of Blackbird, Song Thrush and Mistle Thrush in Great Britain. Pt. 1. Bird Study, 2:2-24. OWEN, J. H The nesting of the Long-tailed Tit. Brit. Birds, 38 :

20 1958 NESTING OF LONG-TAILED TIT 19 PFEIFER, s Die Schwanzmeise als Erdbri ter. Beitr. Fortpfl. biol. V6g., 14: ROBERTSON, A. W. P. & PORTER, S. C Long-tailed Tits' unorthodox nesting arrangements. Brit. Birds, 45: SNOW, D. w The breeding of the Blackbird Turdus merula at Oxford. Ibis, too: 1-3o (espec. p. 25). THIELEMANN, A Schwanzmeise (Aegithalos caudatus) als Bodenbrbter. Ornith. Mitt., 8 :76-7. NESTING HABITS OF THE COLLARED TURTLE DOVE by ZYGMUNT BOCHENSKI (Dept. of Zoopsychology and Ethology, jagiellonian University, Krak6w) Received 14 November SUMMARY The Collared Turtle Dove usually nests on horizontal branches of trees at various heights, though occasional nests are situated on buildings. It is exceptional for it to take over the old nest of another species, e.g. Woodpigeon. The birds take several days to build their nests, but they work only in the morning. The intensity of their work is greater early in the morning than later. Nest building occurs from March to May. INTRODUCTION The Collared Turtle Dove (Streptopelia decaocto) is a bird about which many authors have recently written in European ornithological journals. This is because the boundary line of its geographical distribution is moving northwards very quickly (Miczynski, 195o and 1951 ; Fisher, 1953 ; Stresemann, 195o). The Collared Turtle Dove is a common bird throughout Poland, except the northeast districts (Dyrcz, 1957). Since 195o I have observed many Collared Turtle Doves in Krakow where they live both in the town and in its outskirts, wherever they find trees, which are their typical nest habitat. HABITAT AND NEST-SITE Collared Turtle Doves nest willingly in town gardens and parks. The birds usually build their nests in old trees, on their lower branches at various heights ranging from 15 to 3o ft. above the ground ; Meklenburcev (1951), however, writes that most of the nests he observed were at a height of 6 to to feet. Nests are sited both on horizontal branches near the trunk and also in the outer fringe of the crown. I have one record of a nest in the fork of a horse-chestnut at the height of 20 ft. Exceptionally we can find the nests situated on buildings. Dyrcz (1957) describes three such nests. One of these was built among some ruins ; another in the corner of a gutter and the third on the building at the height of the third storey.