1966 I. NEWTON : THE MOULT OF THE BULLFINCH PYRRHULA PYRRBULA 41
THE MOULT OF THE BULLFINCH PYRRHULA PYRRHULA
I. NEWTON
Received on 13 October 1964
INTRODUCTION Few detailed field studies have yet been carried out on the moult in any bird species. This paper describes some observations on the moult of the Bullfinch Pyrrhula pyrrhula nesa (Carduelinae), and its timing in relation to the breeding season and food-supply in the woodland of Wytham Estate, near Oxford (58" 80'N.). The Bullfinch is a common resident in the area, mainly in the woodland, but also in gardens, orchards, and hedge- rows. It nests in dense shrubs and hedges, and feeds for most of the year mainly on various seeds, and in spring on buds, including those of various fruit trees (Newton 1964, 1964 a). Details of moult and weight were obtained from 133 individuals in 1962, from 283 in 1963, and from 247 in 1964; many of these were handled more than once during a single moult.* They were caught in mist nets placed more or less at random in the woodland or at feeding places such as patches of Meadowsweet Filipendula ulmaria and various docks Rumex spp. Netting was restricted to two areas, the Great Wood and Marley, which have been described by Southern (1954) and Gibb (1954) respectively. The Great Wood comprises some 400 acres of deciduous trees, dominated by Oak Quercus robur, Ash Fraxinus excelsior and Sycamore Acer pseudo-platanus, and about 150 acres of conifers. Marley is an open scrubby area of 66 acres, which is dominated by Oak and Hazel Corylus avellana, with large patches of hawthorns Crataegus spp. and Elder Sambucus niger. Wytham Bullfinches started laying each year in early May, soon after fresh seeds first became available. The laying season lasted until mid-July in 1963 and 1964, but was prolonged to early September in 1962 when food was especially plentiful in late summer. After breeding the adults underwent a complete moult, and at the same time the juveniles underwent a partial moult. The term " moult )' is here taken to cover both the shedding and the renewal of feathers.
Plumages and moults The number and distribution of the feather tracts in the Bullfinch, and the sequence of moults, is similar to those of some other passerines (see Dwight 1900). For the first few days after hatching the nestling Bullfinch is sparsely clothed with natal down. At the post-natal moult into juvenile plumage, each down feather is pushed out by growth of a juvenile feather and clings to the tip of the juvenile feather sometimes until it is fully formed 1-2 weeks after the bird has left the nest. Juvenile body feathers differ from those of the adult in being weaker, looser in texture and of different colour. The post-juvenile moult takes place in late summer or autumn, and is restricted to the body feathers and the small feathers of the wings and tail; the primaries, secondaries, rectrices, the two outer tertials, and primary-coverts, and a variable number of greater-coverts
* Full details of moult were not obtained from every bird, and if time was short details only of wing-moult (or sometimes of wing- and tail-moult) were taken from the adults. This accounts for the different totals in some of the tables. 42 I. NEWTON : THE MOULT OF THE BULLFINCH PYRRHULA PYRRHULA IBIS108 are retained until the next moult about a year later. Birds in first-year plumage can be distinguished by some of the retained juvenile feathers, until at 11-15 months old, the next moult is so advanced that their plumage is indistinguishable from that of older birds. All subsequent moults are complete post-nuptial moults and occur at approxi- mately annual intervals. There are no pre-nuptial moults in the Bullfinch. In an adult Bullfinch, just prior to moult, the feathers are usually worn, faded and soiled, and there are often gaps in the plumage where lost feathers have not been replaced. The fading of the black flight and tail feathers to various shades of brown is presumably caused by the disintegration of some of the cornified cell-layers of the feather (Miller 1928). Fading was marked in birds moulting late, and was most pronounced in birds in first-year plumage in which the flight and tail feathers, acquired as nestlings, could have been up to 15 months old before they were replaced. The adult moult in the Bullfinch lasts 10-12 weeks and the post-juvenile moult 7-9 weeks. In adults replacement of the body plumage rarely spans the entire moult period, but averages about as long as the post-juvenile moult. The moult begins in the adults with the primaries and in the juveniles with the lesser-coverts. The latter are often the first body feathers to be moulted in adult Bullfinches. In this respect the Bullfinch differs from most other passerines in which body moult begins first on the breast, back, scapulars, and head, and only then in the lesser-coverts (Dwight 1900).
THE FEATHER TRACTS AND THEIR PATTERN OF MOULT The distribution of the feather tracts (pterylae) is best seen in a nestling while the feathers are still small. In all there are eight tracts, three of which are paired, arranged as in most other passerines (see Dwight 1900). The moult begins at a number of points (loci) from which it spreads in characteristic manner. There are two loci in the dorsal and ventral tracts, four in the caudal tract, at least five in the alar tracts, and one in the head tract and in each half of the remaining tracts. Normally moult begins at different loci in turn and from each progresses along the length of the tract more rapidly than it spreads laterally.
1. The alar tract The alar tract comprises the flight feathers, their various coverts, and the alula quills. (In addition, the wing bears the scapulars on the upper surface and the axillaries below, both of whch form part of the shoulder tract.) In the Bullfinch, there are nine large primaries, six secondaries, and three tertials. The tenth primary is vestigial as in other carduelines. Moult of the primary feathers begins at the carpal joint and spreads outwards ; the primaries are here numbered 1 to 9, proximal to distal. The secondaries are replaced from the carpal joint inwards, and are here numbered 1 to 6, distal to proximal. Moult of the tertials does not follow the sequence of the secondaries: the central one is shed first, then the innermost, followed by the largest outer tertial. The length of the various flight feathers was measured by pressing the detached quill flat on a rule. The following figures are the means of measurements on 12 birds, and the standard deviation in every case is about & 1 mm. Primaries ' Secondaries Tertials 123456789 123456 123 Length (mm.): 62 63 64 66 71 74 72 69 65 61 61 60 58 56 54 50 40 28
Each primary-covert is shed with its corresponding primary, and has usually reached full length by the time the primary is half grown. The greater-coverts are shed more or less simul- taneously with each other, or in quick succession from the innermost outwards; sometimes however, the inner ones are half grown before the outer four or five are dropped. Replacement of the lesser-coverts begins near the body and spreads first towards the carpal joint and then to the median-coverts, which are moulted all more or less together; at the same time the carpal-covert and the feathers of the alula are replaced. On the underside of the wing the coverts are softer and appear more sparsely distributed; the lesser- and median-coverts are moulted more or less simul- taneously, and the under greater- and under primary-coverts are moulted in rapid succession from the carpal joint outwards.
2. The caudal tract The caudal tract is here considered to include the 12 rectrices, their upper and under coverts and the feathers circling the anus. In the Bullfinch, both upper and under tail-coverts are unusually long, the middle pair of each extending to more than half the length of the rectrices. The rectrices are all about the same length (about 71 mm.)-though in some birds the inner ones are s!ightly 1966 I. NEWTON : THE MOULT OF THE BULLFINCH PYRRHULA PYRRIIULA 43
shorter-and are replaced from the central pair outwards. The ten upper coverts were often shed from the outermost inwards, but this pattern was sometimes modified by the temporary retention of any of them, and in some individuals no orderly sequence of replacement was discerned. The under tail-coverts also tend to be replaced centripetally, though variation was even more frequent than in the upper coverts. The feathers circling the anus are all moulted together.
3. The dorsal tract The dorsal tract includes the feathers of the nape, back, and rump. It is continuous with the head tract, but posteriorly, stops in front of the preen gland. Feather replacement begins at two loci, one between the scapulars and the other just above the rump and spreads both ways along the length of the tract, the nape being the last part to acquire new feathers.
4. The ventral tract The ventral tract starts under the bill and about half way down the neck forks into two broad symmetrical bands; each in turn divides to give a short outer branch in front of the leg, and a main branch extending down the belly to end in front of the anus. Feather replacement begins at two loci, one in each of the main branches, and spreads lengthwise along the tract. Progress is more rapid forwards, and the feathers of the belly are among the last to be replaced. This tract comprises all the ventral plumage except the feathers of the flanks and thighs and the under tail-coverts. 5. The head tract This tract covers the entire head and passes into the dorsal tract behind. Moult begins round the beak and passes back over the crown.
6. The thigh tract The thigh tract comprises two narrow oblique bands of feathers, one on each flank. Moult begins at the front and spreads back. 7. The shoulder tract Each half of the shoulder tract connects with the ventral tract, then curves round dorsally to form a narrow band of feathers (scapulars) across the humerus. Feather replacement begins at the front end and spreads back through the scapulars to the tertials.
8. The mural tract The crural tract comprises the feathers of the legs.
THE SEQUENCE OF MOULT IN ADULTS THE METHOD USED TO RECORD THE STATE OF ADULT MOULT Moult begins with the first primary and in most birds the replacement of the last one marks its end. The stage of primary moult provides a fairly good indication of the stage of moult in all the other feathers. The simple numerical system, previously used by Miller (1961) and Ashmole (1962) and based on the moult of the primaries, was therefore used to record the state of moult in adult Bullfinches. Each primary of each wing was given a score from 0 (old) to 5 (new) according to its state of growth. The sum of the scores of individual primaries (the " primary score ") was then taken as an index of the general state of moult. Since the Bullfinch has nine primaries in each wing, a fully moulted bird would have a primary score of 90. In recording the state of moult of individual feathers, the following categories and scores were used:- Old feather 0 Feather missing or in small pin stage 1 Feather in large pin or brush stage 2 Feather brush to half-grown 3 Feather half to three-quarters grown 4 Feather three-quarters to full length 5 For birds which were handled more than once during moult, it was sometimes possible to work out the growth rate of individual feathers; and as far as could be ascertained, each of the above stages usually took an equal part of the total time required for the feather to grow (see later). 44 I. NEWTON : THE MOULT OF THE BULLFINCH PYBRHULA PYRRHULA IBIS108
THE PATTERN OF MOULT IN THE PRIMARIES AND SECONDARIES As mentioned above, the shedding of the first primary signified the onset of moult, and when this was partly grown, the second was shed and so on through the series. Replacement of the secondaries started about half way through the primary moult, by which time the inner two or three primaries had reached full length. Corresponding feathers in each wing were usually shed on the same day, but sometimes, a feather in one wing was as much as half grown before its equivalent in the other wing was dropped. The moult in adult Bullfinches was divided into nine successive stages based on the primary score. Stage 1 included birds with scores 1-10, stage 2 those with scores 11-20, and so on. Table 1 gives the mean number of growing and new primaries and secondaries for birds at different stages of moult, together with details on the moult of individual feathers. For a time at the start and end of moult, there was only one growing primary in each wing, but through most of the moult there were usually two or three. There was however considerable individual variation, especially about the middle of moult, when some Bullfinches had up to six growing primaries in each wing. It was noticed also that birds moulting particularly late in the season often had more primaries growing simultaneously than those moulting earlier. The first secondary was shed as the fifth primary was moulted. One third of the birds at stage 4, and all the birds at later stages of moult had at least one growing secondary in each wing. Replacement of the first three secondaries proceeded slowly but the last three were shed in quick succession, so that many birds at stage 9 had three or four secondaries growing concurrently. The rate of increase in secondary score was therefore slow at first, but increased abruptly towards the end of moult (Fig. 1). Progressive acceleration in the rate of secondary moult is apparently general in passerines (Dwight 1900). In about a third of the Bullfinches examined, primary moult ended after secondary moult; in about a half, primary moult ended a few days before secondary moult; and in the rest, moult of the two series finished together.
secondary score
60. -: 3= 50- e
0. *d 40- C.
4- 30- :** 20- 0. ** 10- *. I* ** a**** :**-** ** ~-ooooo*wa**w*moe ***:* 1 ..
TABLE1. The moult of the primary and secondary feathers in the Bullfinch Pyrrhula pyrrhula. Mean Mean Mean Mean No. of birds with the followin8 feathers in growth:- no: total no. no. growing no: filll- full- No. of Primaries Secondaries second- growing Fowp grown Moult aria renuges primma second- stage %zexzed 123 4 5 6 7 8 9 123 4 5 aries ez 1 1-10 44 4420 J 1.6 m 2 11-20 25 22 25 20 I 3 .O 0.08 0.08 m 3 21-30 21 9 17 25 19 4 2.8 1.1 1.1 c 4 314 11 1 3 13 11 16 1 4 0.2 3.2 2.4 5 41-50 14 4 12 14 10 12 1 1*o 3.8 2.9 6 51-60 24 3 11624 9 1 224 1.4 4.2 4.2 0.8 I 61-70 12 191192 67222 1*8 4.6 5.2 0.4 8 71-80 8 788 7741 2.5 5.4 6.4 1.2 9 81-89 31 2 15 31 IS 17 23 2 4 4.5 7.3 3.2 Note. For explanation of atago of moult and primary nmm sea text. 46 I. NEWTON : THE MOULT OF THE BULLFINCH PYRRHULA PYRRHULA IBIS108
THE MOULT OF THE RECTRICES All the 12 rectrices are about as long as the longest primary. In recording the growth-stage of individual feathers, the same notation was used as for the flight feathers. Most of the 84 birds examined (Table 2) which were moulting tail feathers moulted them symmetrically, the second pair being shed when the first had reached pin stage, and so on until five or six pairs were simultaneously in moult. In some cases, however, adjacent pairs of feathers were shed in quick succession so that the eight central feathers were at pin stage simultaneously, though each pair was of slightly different length. In other cases the feathers had been dropped more slowly, so that the middle three pairs had reached full length before the outer two pairs were shed. In 11 of the Bullfinches examined, moult of the rectrices was not symmetrical, either because some of the feathers had been accidentally lost prematurely or because feathers had been lost faster on one side of the tail than on the other. In no case was the disruption of sequence so great as to obscure the underlying centrifugal plan of moult. TABLE2. The moult of the rectrices in the Bullfinch Pyrrhula pyrrhula No. of birds with following Primary No. of No. with pairs of rectrices in moult Mean no. of moult birds tail growing rectrices in stage examined moult 1 2 3 4 5 6 birds with tail moult 1-2 64 0 3 16 7 761 2-5 4 16 13 13 12 7 1 5.1 5 14 14 14 14 13 9 4 6.5 6 22 22 19 20 19 22 15 10 10 7 12 12 3 3 9 91111 5.2 8 8 8 48 3 9 31 8 8 2 Note. For explanation of stages of primary moult see text. Replacement of the rectrices normally began as the 4th or 5th primary was moulted, and continued until the 9th was in course of replacement. Most of the birds examined at stage 4, and all birds at stages 5 to 8, had growing rectrices (Table 2). The score of the rectrices increased fastest in stage 6 when all, or almost all, the rectrices were growing concurrently (Fig. 2).
Tail score
50- : . . .0. . Oi .:*: 30- 0%
20-
10-
10 20 30 40 50 60 70 80 90 Primary score
FIGURE2. Tail score in relation to primary score in moulting adult Bullfinches Pyrrhula pyrrhula. Tail score (0-60) and primary score (0-90), which are calculated from the growth stage of individual feathers (see text), indicate the state of moult in the rectrices and primaries respectively. 1966 I. NEWTON : THE MOULT OF THE BULLFINCH PYRREULA PYRRHULA 47
The growth rate of individual rectrices could be worked out for only three birds. In the first, two feathers had each gained one point in seven days (2 mm. per day); in the second, four feathers had each gained one point in eight days (1.8 mm. per day); and in the third, five feathers had each gained one point in four days (3.5 mm. per day). These figures show considerable individual variation but that the growth rate of the rectrices can be as fast as that of the remiges, which averaged 3-4 mm. per day (see later).
THE MOULT OF THE REMAINING FEATHERS In table 3 the timing and duration of moult in all other tracts, and the remaining parts of the alar and caudal tracts, are shown in relation to primary moult: the figures show the number of birds examined at each stage of moult which were then moulting particular feathers.
TABLE3. The moult of the body feathers, and of the small feathers of the alar and caudal tracts, of the Bullfinch Pyrrhula pyrrhula in relation to primary moult. The jigures show the number of bid examined at each stage of moult which were moulting particular feather tracts. For explanalion of stages of moult see text. Stage of primary moult 1 2 3 4 5 6 7 8 9 Total number of birds examined : 40 24 20 14 13 20 10 8 31 Feather group Upper greater-coverts 4 14 14 8 1 median-coverts 13 9 8 10 5 lesser-coverts 9 14 17 11 7 11 4 Under greater-converts 4 5 6 13 median-coverts 4 8 7 6 2 lesser-coverts 6 7 13 8 5 2 Tertials 12 9 12 16 9 2 Alula 1 2 1 1 2 5 5 Scapulars (shoulder tract) 6 15 16 11 8 10 6 2 13 Axillaries 2 10 14 6 2 Ventral tract 8 16 16 14 13 20 10 8 21 Dorsal tract 9 13 17 14 13 20 10 5 15 Thigh tract 3 9 8 11 17 9 8 27 Leg 5 10 9 13 17 4 4 6 Head 4 9 13 i3 20 8 5 6 Upper tail-coverts 2 12 11 17 10 5 12 Under tail-coverts 6 13 13 17 9 4 7
(1) Birds without dependent young In a bird without dependent young (see later), moult of the body feathers, and of the small feathers of the alar tract, usually began with the moult of the second or third primary feather. (In at least three birds however, body moult and primary moult began simultaneously.) The upper lesser-coverts were the first to begin moult, followed by the feathers of the dorsal, ventral and shoulder tracts, and then the upper greater- coverts. In these independent birds, this order of inception of moult at different moult loci was seldom varied. Moult in the leg, thigh and head tracts began soon afterwards, followed by the lesser-coverts of the underwing, the middle tertial, the rectrices and their coverts. Many of the birds examined at stages 4-6 showed moult in every group of feathers, though the greater-coverts of the underwing were usually replaced later at stages 7-9. The upper greater-coverts had been completely replaced in all birds by stage 7, the upper lesser- and median-coverts by stage 8, and the under lesser- and median-coverts, and the tertials by stage 9. The feathers of the ventral, dorsal, head, thigh and leg tracts, together with the tail-coverts were still being replaced almost at the end of primary 48 I. NEWTON : THE MOULT OF THE BULLFINCH PYRRHULA PYRREULA IBIS108 moult. However, there was considerable variation and after stage 6 most birds were clothed almost entirely with new and growing feathers, and appeared fully moulted when seen in the field. Thus the most intense renewal of body feathers occurred at stages 3-6, when almost every tract and group of feathers was in heavy moult. But full replacement of the plumage of the dorsal and ventral tracts, two of the largest tracts on the body, lasted almost the entire moult period. Feathers on the belly and head were the last to be moulted.
(2) Bivds with dependent young The start of body-moult in relation to primary moult showed considerable individual variation, but most of the birds with delayed body-moult were known to have dependent young (see later), and at first showed a different sequence of moult to those without young. Not only were the primaries shed more slowly (see later), but regeneration of new body feathers was considerably delayed, even though large areas round several moult loci had become bare of feathers. Then, presumably because the young reached independence, growth started abruptly and simultaneously in several tracts and spread so rapidly that the whole bird appeared covered with growing feathers. Such birds provided a striking contrast to those apparently without dependent young, in which body- moult began in sequence in each tract, and proceeded at a fairly constant rate. This agrees with Michener & Michener (1940) who found that in those House Finches Curpoducus mexicunus which began to moult before mid-summer, there was a considerable interval (a third of the total moult period in one bird) between the start of moult in the primaries and their coverts on the one hand, and in the rest of the plumage on the other. In later-moulting adults, this delay was reduced or absent. Also Miller (1961) found that in the Andean Sparrow Zonotrichiu cupensis, the timing of events within the moult cycle varied greatly: in some individuals the body feathers were the first to start moulting, but in others body-moult did not start until the 3rd or 4th primary had been dropped. The delay in the start of body-moult in some individuals of both these American species could have been related, as in the Bullfinch, to the encroachment of breeding activity on the normal moult period.
THE RATE, DURATION AND TIMING OF ADULT MOULT
METHODS OF ESTIMATING THE RATE, DURATION AND TIMING OF ADULT MOULT An ideal system of scoring the stage of moult would ensure that equal increments in primary score would connote equal periods of time. This would occur if, for instance, successive primaries were shed at regular intervals, and if each primary, irrespective of its length, took the same time to grow. Information on the rate of increase in primary score for individual birds was obtained from those caught more than once during moult, while estimates of the mean rate, duration and timing of moult were obtained from regression analyses of primary score on date. The data for the two later years were divided into four groups, those for yearling and older males, and yearling and older females respectively (Table 4). In each case, a regression analysis of primary score on date showed a straight-line relationship, the linearity of which was tested by examining the departure from linearity of samples of more than one primary score on each day. It is perhaps surprising that a linear relationship should exist between mean primary score and date in the Bullfinch, since for any one individual, a plot of primary score against date might be expected to give a curve of sigmoid shape, the primary score increasing more slowly at the start and end of moult when fewer than average feathers were growing (Table 1). However, the composite scatter diagrams (Fig. 3), obtained 1966 I. NEWTON : THE MOULT OF THE BULLFINCH PYRRHULA PYRRHULA 49
TABLE4. Estimaies .f the mean rate, duration and timing of moult in adult Bul&nches Pyrrhula pyrrhula from regression analyses.
Yearling 6 Older 6 Yearling 9 Older 0 1963 Sample size* 37 22 33 7 Regression equation P==-0.6+ 1'104D P== - 10.29fl~160D P=-8.19+1.19ID P--l2.96+1.23D Rate of increase in primary score (points per day) 1.104*0.09 1.16 1.0.13 1.191 50.10 1.229 f0.19 Duration of moult (days) 8 I .58 77.57 15.56 73.19 Start of moult Aug.,lf 10.3 Aug,8f10.3 Aug;Jf 10.7 Aug. 11 :t 10.2 Linearity test FL4-11-0.61 I?':= 1.17 F;Y= 1.53 -
1964 Samnle she* 24 27 20 17 Re-&&Gequation P=1.22+1.095D P= -12.33+1.18D P= -4.29+1.123D P- - 10.11+ 1.133D Rate of increase in primary score (points per day) 1.095*0.19 1.18 & 0.059 1.123 f0.088 1.133j~0.14 Duration of moult (days) 82.17 76.26 80.15 79.43 Start of moult Aug. 1 i14.9 Aug. 10f7.3 Aug. 9;tg.l Aug. 9k8.4 Linearity test Fi =243 F1:= 0.589 F1:=2.24 P1g=0.88
Note. Includes retraps, but unusually early or late individuals were omitted from analysis since these can bias the estimates obtained, especially if the samples are small and the data are not collected regularly throughout the moult period. by plotting the primary scores of many individuals against date, show little trace of a sigmoid shape, probably because in the Bullfinch the period when fewer than average feathers are growing is short compared with the spread in the date of onset of moult in the whole population. It is worth mentioning that since the different age and sex groups began moult, on average, on different dates, if the data for all birds are combined, the spread in the start of moult is then so great that a regression analysis gives a value for the overall rate of moult which is too small. Such an analysis can therefore be used to estimate the rate of moult satisfactorily only when the time of moult in a population is fairly well synchronised-otherwise the regression analysis gives niisleading results. (For further discussion of the use of regression analyses on moult data see Evans, in press.)
THE RATE AND DURATION OF ADULT MOULT The mean rate of increase in primary score determined by the regression analyses varied from about 1.10 to 1.23 points per day, and the duration of moult from 73 to 82 days (Table 5). In both years the yearling males examined started moult on average a few days earlier than the other groups, probably because some of them were non-breeding birds (Newton 1964). Information on the rate of moult in individual birds is available for 21 birds which were handled more than once during a moult, and at intervals up to 68 days. The rate of increase in primary score in some of these birds is shown in Fig. 4 in which each line refers to one bird, and its slope is the rate of moult in points per day. The mean rate of increase in primary score was 1.30 & 0.2 points per day, which is consistent with the estimates obtained from the regression analyses. Recorded over shorter periods, the slowest rate of progress was 0.4 points per day over 14 days, and the fastest six times as great, 2.5 points per day over 15 days. Variations in the rate of primary moult in three individuals caught more than twice during moult are shown in Table 5. The fastest rate of primary moult recorded in bird B1, was almost three times faster than the slowest. In B2, the rate at which the primary score increased varied from 1.33 to 2.25 points per day; while B3 maintained about the same rate throughout. The greatest deviations from the mean rate of moult D VOL. 108 50 I. NEWI'ON : THE MOULT OF THE BULLFINCH PYRRHOLA PYRRHULA IBIS108 Primary score
0 I0I30 .. O. 0 1962
10- 0
1
30 .... .O . 0 0. . 1963 8.. & to 10- 0 mp OO r:. 00 - 0 0.
August Septern ber October November December
Primary score
. 0 . .:! 0 . 0 O. 00
0 . 30- 0 ..*: 0. 0 0 1964 . .* 10- .. . 8 00 0 :. 0. 0 .. 1966 I. NEWTON : THE MOULT OF THE BULLFINCH PYRRHULA PYRRHULA 51 appeared in birds in which the interval between successive observations was short. Thus the longest lines in Fig. 4 run nearly parallel, indicating that the overall rate of moult was about the same in all these birds, even though the rate may have varied widely at different stages of moult. Primary score
August Sopt e m ber October FIGURE4. The rate of increase in primary score in relation to date in adult Bullfinches Pyrrhula pyrrhulu. Each line refers to a particular individual, and its slope indicates the rate of moult in the primary feathers in points per day. (For explanation of primary score see text.)
TABLE5. Fluctuations in the rate of moult in three adult BullJinches Pyrrhula pyrrhula.
Rate of increase in Date of moult score in points Bird capture Score per day B1 1 Aug. 63 4 1-46 (11 days) 12 Aug. 63 20 1.77 (26 days) 7 Sept. 63 66 0.67 (3 days) 10 Sept. 63 68 0.89 (18 days) 28 Sept. 63 84
B2. 15Aug. 63 8 1.43 (7 days) 22 Aug. 63 18 2.25 (16 days) 7 Sept. 63 54 1.33 (9 days) 16 Sept. 63 66
B3 10Aug. 63 8 1.50 (10 days) 20 Aug. 63 23 1.43 (14 days) 3 Sept. 63 43 1.48 (21 days) 24 Sept. 63 74 Note. The number of days between successive observations is given in parenthesis in column 3. 52 I. NEWTON : THE MOULT OF THE BULLFINCH PYRRHULA PYRRHU~LA IBIS108
Assuming a mean rate of 1.3 & 0.2 points per day, the time taken to moult would be 69 days (+-9 days), plus up to about seven more days in those individuals in which moult of the secondaries ended after moult of the primaries. Three individuals were caught soon after the start and again just before the end of moult, and assuming that each primary grows at a rate of 0.4 points per day (see later), the time taken for the complete moult in these birds would be 68, 70, and 74 days respectively, estimates which are again consistent with those obtained from the regression analyses (Table 4). The duration of moult in adult Bullfinches is similar to that of some other passerines of temperate regions, such as Steller’s Jay Cyanocitta stelleri (Pitelka 1958). On the other hand, Lesser Redpolls Curduelis$ummeu cabaret moult in only eight weeks (Evans, in press) and Chaffinches in only six (Marler 1956). In the House Finch of California moult lasts about 15 weeks (Michener & Michener 1940), and in the Boat-tailed Grackle 13 weeks in the female, compared with 15 in the male (Selander 1958).
THE CAUSE OF VARIATIONS IN THE RATE OF ADULT MOULT Variations in the rate of increase in primary score in the Bullfinch could have been due to differences in the growth rate of individual feathers, or to variations in the number of feathers growing simultaneously, or to the maintenance of the same growth rate in feathers of different lengths. (The last alternative would result in a slightly slower increase in primary score during the second half of moult, when the longer outer feathers were being replaced, since the same scoring system was used for each feather irrespective of its length.) In addition, the rate of any of these processes could have been influenced by the physiological state of the bird, or by environmental factors such as the availability of food. Information on the growth rates of individual feathers is available for only six birds and covers primaries 1 to 6, which vary in length when full grown from 62 mm. to 74 mm., the shortest and longest respectively. Growth data were also obtained for the first secondary (62 mm.). In all six birds, all these feathers grew so as to gain one point in four days (3-4 mm. per day) even though the number of growing primaries in each wing varied in different birds from one to four. There was therefore little or no variation in the time taken for these particular feathers to grow in the six birds examined, implying that the longer feathers grew slightly faster than the short ones, and that full growth of any one of them took about 20 days or just over. The rate at which the primary score increases will also depend on the number of growing feathers involved. The mean number of growing primaries in each wing remained at 2-8-3.0 throughout the moult, except at the start and end when one primary grew alone (Table 1). There was however considerable individual variation, and some birds had up to six growing primaries in each wing, though this was rare. Of 21 birds caught more than once during moult, the interval between the shedding of successive primaries varied from 5 to 10 days, with an average of 7 days, as would be expected from the duration of moult. There is evidence that food-shortage slows the moult in certain species. When Riddle (1908) began starving doves ten days after plucking certain feathers, the linear growth rate of the feathers was at first.maintained, but dropped three days later. Ashmole (1962) found that in a population of the Black Noddy Anous tenuirostris primary moult was slower when food was inferred (on other grounds) to be scarce than when it became more plentiful. Although most Bullfinches moulted while food was still abundant (see later), those birds feeding young until comparatively late in the moult (to stage 3) shed their primaries more slowly than birds apparently free of dependent young. Table 6 shows that individuals with young (see later) had one or two primaries growing simul- taneously in each wing, one of which had usually reached almost full-length before the 1966 I. NEWTON : THE MOULT OF THE BULLFINCH PYRRHULA PPRRHULA 53
next was dropped; while birds apparently free of young, but with the same moult score, had up to five primaries growing in each wing and never less than two. Birds at earlier stages of moult showed no such difference.
TABLE6. The pattern of wing-moult in adult BullJinches Pyrrhula pyrrhula with dependent young compared to that in birds apparently free of young.
Individual Score of urimaries:- Total number of growing primaries 123456 per wing Primary score 20-22 ---- With young 1 (male) 55 a}l.5 2 (female) 5 4 I--- Without young 3 (male) 542--- 4 (male) 333 2-- :>, 5 (female) 5 3 2 1-- 3 Primary score 28-30 With young 6 (male) 555--- 7 (male) 555--- ;}*.3 8 (female) 5 5 4 1-- 2 Without young 9 (male) 55 3 I-- 10 (male) 55 2 2-- 11 (male) 433 2 2- 12 (female) 5 4 3 2 1- 13 (female) 5 5 4 I-- 14 (female) 5 4 321- Note. For explanation of primary score see text.
This apparent slowing in the shedding of primaries in birds with young could have been directly due to " shortage " of food imposed by the dual strain of moult and parental care. Although there are insufficient data to compare the growth rates of individual feathers in birds of each category, " fault bars " were never seen on the feathers of any Bullfinches examined. These are areas of weakness produced on the feathers of some species when food is short and appear as pale bars across the vane where the barbs are deficient in barbules and melanin (Riddle 1908). To sum up, in the Bullfinch different primaries appear to take about the same time to complete their growth, though the data on this point are limited. Differences in the rate of moult between individuals, and at different stages of moult in the same individual, are due mainly to variations in the numbers of growing feathers involved. In addition, there is some slight evidence that moult in the Bullfinch may be influenced by other factors, since birds which were feeding young dropped their primaries more slowly than those which were not.
THE TIMING OF ADULT MOULT Fig. 5 shows the number of birds beginning to moult in five-day periods throughout each of the three years. The date that each bird started to moult was estimated by working back at 1.3 points per day from its score on the date of capture (Fig. 3), except for those birds caught at the beginning of moult with only the first primary growing, when a rate of 0.4 points per day has been assumed (see above). Most birds were caught near the beginning of moult, and few estimates should be in error by more than one five-day period. In 1962 the start of moult in the adults was spread over the 11 weeks from the end of July to the beginning of October, and in the following two years over the six weeks from the end of July to the beginning of September. The greater spread in time of onset of moult in 1962 is reflected particularly well in the range of primary scores of birds 54 I. NEWTON : THE MOULT OF THE BULLFINCH I’YRRHULA PYRRIIULA IBIS 108
25-
20- 1962 1963 1964 Total number Total number of Total number of oi birds examined = 22 birds oxaminud=96 birds oxaminedz 83
15-
10-
5-
I _Ia
caught at the end of September. For instance, on 24 September 1962, six adults were caught, one of which had just started to moult (score 6), and another (score 84) must have started nine or ten weeks earlier. The mean date of onset of moult in 1962 was 31 August (S.D. 23 days), compared with 9 August (S.D. 11 days) in 1963 and 11 August (S.D. 12 days) in 1964. In all years, the adults which moulted earliest finished in early October; in 1962 there were moulting adults in the population almost until the end of December, but only until mid-November in the two following years.
Percentage in moult 1962 1963
Percentage in mouIt 1964 100
50
0
FIGURE6. The moulting season of Bullfinches Pyrrhulu pyrrhtilu near Oxford in three successive years, Continuous line=adults; broken line =juveniles. 1966 I. NEWTON : THE MOULT OF THE BULLFINCH PYRREULA PYRREULA 55
Ten birds were caught during moult in two different years, and most of them moulted at a different time (with respect to the mean date of moult) in each year (Table 7). These differences presumably result from variations in the time they finished breeding in different years (see below).
TABLE7. The date of onset of moult in individual BullJinches Pyrrhula pyrrhula in different years. The jigures in parentheses show the number of days before (-) or after (+) the mean date of onset of moult in the population as a whole.
Estimated date of start of moult 1962 1963 1964 Mean date 31 Aug. 9 Aug. 11 Aug. First-year* males - 28 JuI. (-12) 18 Au~.($9) - 7 Aug. (-2) 9 Aug. (-2) Adult* males - 6 Aug. (- 3) 20 Aug. (+9) - 1 Aug. (- 8) 6 Aug. (- 5)
First-year* females 9 Sep. (+ 9) - 18 Aug. (+7) - 2 Aug. (-7) 3 Sep. (f23) - 27 Jul. (-13) 3 Aug. (-8) - 4 Aug. (-5) 20 Aug. ($9) - 28 Aug. (+19) 19 Aug. (+8) Adult* female - 6 Aug. (-3) 16 Aug. (+5) * Age-group when first examined.
THE RELATIONSHIP BETWEEN THE END OF BREEDING AND THE ONSET OF MOULT Breeding and moult are processes of high energy demand, and in almost all species of temperate regions occur in summer when conditions are favourable (Pitelka 1958). In many species, including the Bullfinch, the moult of the adults begins as their last young become independent. In all three years Bullfinches started nesting in early May, but this initial synchrony was soon lost, mainly through the action of predators (Newton 1964). Presumably it was the subsequent lack of synchrony in breeding activity which caused the spread in the time of onset of moult in the population, since each pair started to moult as their last young left the nest. The greater spread in the time of onsetof moult in 1962 than in the two following years was due to a proportion of pairs having additional broods begun after the end of July (Newton 1964). Such late breeding occurred in response to a comparative abundance of food in late summer (see later and Newton 1964), and the onset of moult was thereby delayed.
TABLE8. The number of adult BullJinches Pyrrhula pyrrhula at different stages of moult which were known to have dependent young.
Stage of moult Moult not Stage 1 Stage 2 Stage 3 Stage 4 begun Total number of Bullfinches caught 44 44 25 27 17 Total number feeding young 16* 15* 5* 5* 0 *Note. Birds were known to have dependent young only if they were seen feeding them, or if they had food in the throat pouch. The figures, which are based on all adults caught August- October all years, are therefore minimum figures.
In most Bullfinches parental care and moult tend to be mutually exclusive, but in all three years many of the adults caught in the early stages of moult still had dependent young, some of which were seen being fed; while some adults had food in the throat 56 I. NEWTON : THE MOULT OF THE BULLFINCH PYRRHULA PYRREULA IBIS 108 pouch when caught. Young Bullfinches are normally fed by their parents for 1-2 weeks after fledging. The number of adults at different stages of moult which were known to have dependent young is shown in Table 8, which is based on the data for August to October in all three years. At least a third of all the birds caught at stage 1 of moult, and a fifth of those at stages 2 and 3 had dependent young, but none at any later stage. In those at stage 3 the moult presumably began before their young had left the nest, unless the young were dependent for an unusually long period. The overlap between moult and dependence of young occurred not only in those birds in which the onset of moult was delayed by late breeding, but also in those starting to moult comparatively early. The earliest breeding Bullfinch in moult was a male caught on 6 August 1963 (score 4). On five occasions both members of known pairs (caught earlier at their nests) were caught, and in each case the male’s moult was slightly ahead of the female’s. Similar slight overlap between parental care and moult has been noted in some individuals of at least four other passerines of temperate regions, the Northern Water- thrush Seiurus noveboracensis (Eaton 1957), the Plain Tit Parus inornutus (Dixon 1962), the Lesser Redpoll Carduelis fiammeu cabaret (Evans, in press) and the Chaffinch FringiZZu coelebs (Marler 1956). Experimental evidence is available for several species, such as the Red-winged Blackbird (Wright & Wright 1944) and the White-crowned Sparrow Zonotrichia Zeucophris (Blanchard 1941), that the moult in the males begins after the testes have regressed so far that no free sperms are present. In the male Andean Sparrow, however, the correlation between testis regression and moult is only partial, though in the female of this species breeding and moult are mutually exclusive (Miller 1961). Some larger species of birds, on the other hand, show a protracted replacement of the remiges which begins during the breeding season; this applies for instance to the Accipitridae, which depend on full wing-efficiency for their livelihood (Heinroth 1931). In addition, many species, such as the Glaucous Gull Lams hyperboreus (Johnston 1961), breed and moult concurrently during the short arctic summer.
THE POST-JUVENILE MOULT The post-juvenile moult was limited to the body feathers and the small feathers of the alar and caudal tracts; these feathers comprise about three-quarters of the weight of plumage replaced by the adults (see later). The sequence of post-juvenile moult was in most respects similar to the body moult of adults which had no dependent young. Table 9 shows some of the feathers retained by juvenile male and female Bullfinches according to the date at which they started moulting. The figures are based on all juveniles caught in 1962 and 1963 after completing a moult of known date. In addition to the primaries, secondaries, two outer tertials, rectrices, primary-coverts and alula, all juveniles retained the alula-covert, the carpal-covert, or both. Many birds moulting late in the season retained up to five greater-coverts; only those birds which moulted earliest replaced them all. Late-moulting juveniles were never noticed to have the under wing-coverts in moult. They were among the last feathers acquired by the juvenile after leaving the nest, and they too were probably retained by birds moulting late in the season. Exceptionally a few body feathers were also retained, usually in areas that are normally late to moult, such as the belly. Table 9 shows that juvenile males always replaced more plumage than juvenile females moulting at the same time, but the reasons for this are not known. The retention of the juvenile alula-covert or carpal-covert, which are edged with brown, distinguished first-year birds from adults, in which they are edged with grey. Since only certain feathers are replaced at the post-juvenile moult, presumably such replacements are necessary to reinforce the most vulnerable parts of the plumage; although not seriously damaged by the time of post-juvenile moult, they might, if not replaced, become dangerously worn by the next moult. The strongest and stiffest 1966 I. NEWTON : THE MOULT OF THE BULLFINCH PYRRHULA PYRRHULA 57
TABLE9. The number of juvenile feathers retained by juvenile Bullfinches Pyrrhula pyrrhula moulting at different times. Number of birds which retained the following feathers :- Moult began Sex No. of birds alula carpal greater-coverts* examined coverts coverts 1 2 3 4 5 July Male 3 2 3 00000 Female 3 3 3 11000 August Male 34 34 33 44200 Female 23 22 23 14 14 6 4 1 September Male 25 25 25 18 18 14 8 0 Female 20 20 20 20 20 16 12 4 October Male 8 8 8 7753 Female 6 6 6 66666 * Numbered from the outside inwards, feathers, the flight and tail feathers and primary-coverts, were always retained; the less resilient greater- and alula-coverts were renewed in all but late-hatched young; while the lax body-feathers were replaced by all birds. The retention of those feathers most resistant to wear appears to be fairly general in passerines (Dwight 1900), but there are exceptions such as the House Sparrow Passer domesticus in which the juveniles replace all their feathers at the post-juvenile moult. Bullfinches hatched late in the season generally began to moult within a fortnight of leaving the nest (see later), when they had only just acquired juvenile plumage, or even while parts of it, such as the rectrices and under wing-coverts, were still growing. They would presumably gain nothing by replacing feathers only recently acquired. The energy saved by the late-hatched young in retaining feathers that are replaced by older young is, however, negligible, since the feathers involved weigh less than a hundredth of those that are moulted (see later). A tendency of late-hatched young to replace fewer juvenile feathers than those hatched earlier has been noted also in the House Finch (Jlichener & Michener 1940), Boat-tailed Grackle Cassidix mexicanus prosopidicoZa (Selander 1958), Phainopepla Phainopepla nitens (Miller 1933), Brown-headed Cowbird Molothrus ater and Red-winged Blackbird AgeZaius phoeniceus (Selander & Giller 1960).
THE METHOD USED TO RECORD THE STATE OF POST-JUVENILE MOULT The primaries, used to record the state of adult moult, are not replaced at post- juvenile moult. However, five distinct stages of body-moult, each of the same duration, were recognised as follows :- Stage 1. Moult has begun in the upper lesser-coverts (but not in the upper greater- coverts) and usually on the shoulder and breast, and sometimes on the back and leg as well. Stage 2. Moult has begun in the greater-coverts but not in the thigh tract. A few birds at this stage have also begun to moult on the forehead. Stage 3. Moult has begun in the lower tail-coverts, upper median-coverts, thigh tract and head, but the greater-coverts have not yet reached full length. Also, moult of the upper lesser-coverts may be almost completed, and in a few birds, moult of the upper tail-coverts has started. Stage 4. Both upper and lower tail-coverts in moult. The upper lesser- and grearer- coverts are full grown, the moult of the head feathers has reached the crown, and the moult of the shoulder tract is almost complete. Stage 5. Brown juvenile feathers remain in a small patch on the belly, and also a few in most other tracts, especially on the head. 58 I. NEWTON : THE MOULT OF THE BULLFINCH PPRKHULA PYRRHULA IBIS108
THE RATE AND DURATION OF POST-JUVENILE MOULT Information on the rate of post-juvenile moult was obtained from nine birds in 1962 which were caught more than once during the moult, from 39 in 1963, and from 27 in 1964, covering all stages of moult. The rate of moult was measured as the number of stages completed between successive observations, but the limitations of the recording technique make it possible only to give rough measurements of the rate of post-juvenile moult and the extent of individual variation. The retrap data show that in most birds each of the five stages lasted 9-12 days. Hence, while adult moult lasts about 10-12 weeks, post-juvenile moult normally lasts only 7-9 weeks. Two birds which had not begun to moult when first caught were retrapped fully moulted 56 and 60 days later respectively, while a third took 53 days to complete the first five stages of moult. Different rates of moult were observed in only six individuals: the quickest was 22 days from stage 1 to 5-about twice the normal rate-and another juvenile, which had not started to moult on 2 September, had reached stage 5 a month later; the slowest rates were recorded in three birds which remained at stage 1 for at least a fortnight. Thus, fairly constant rates of progress were found in most juveniles that were handled more than once during moult, but the few discrepancies found were large. There is little comparative information available for other species, but in the Andean Sparrow both adult and post-juvenile moults took about the same time to complete, though as in the Bullfinch there was greater variation in the duration of post-juvenile moult than in that of adult moult (Miller 1961). In the Loggerhead Shrike the post- juvenile moult, which is partial, lasts 36 months compared with three months for the adult moult, which is complete (Miller 1928).
THE TIMING OF POST-JUVENILE MOULT Information on the start of moult is available for 111 juveniles caught in 1962, for 187 in 1963, and for 164 in 1964 (Table 10). For each individual caught in moult, the approximate date of onset was estimated from the state of moult at the time of capture, assuming that each stage lasted 9-12 days. Juveniles caught after the end of July, but which were not moulting, were assumed to begin moult within ten days of the date of capture (see later). Since the method of recording post-juvenile moult was less precise than that used for adults the data have been grouped into 10-day periods. Few juveniles moulted at a rate which differed markedly from the mean rate (see above), so the estimates are probably reasonably accurate for most birds, and few should be in error by more than one 10-day period.
TABLE10. The percentage of juvenile Bulljinches Pyrrhula pyrrhula starting to moult in each ten-day period over three successive years.
Ten-day periods Total Year caught July August September October November 3 123123123123 1962 111 3 9 12 17 15 13 5 10 13 2 1 0 0 1963 187 13 29161916 6 100 0 0 0 0 1964 164 12 26132313 7 6 0 0 0 0 0 0
In 1962, the onset of moult among the juveniles was spread fairly evenly over the 14 weeks from the end of July to the beginning of November (Table 10) and some were moulting to nearly the end of December (Fig. 6). In the following two years, the onset was spread over only nine weeks, from the end of July to the end of September, and there were no moulting juveniles in the population after about mid-November (Fig. 6). 1966 I. NEWTON : THE MOULT OF THE BULLFINCH PPRRHDLA PYRRHULA 59
In all years there was a period when juveniles that had completed their moult occurred in the population simultaneously with juveniles in which the moult had not begun. In 1962 the overlap lasted 6-7 weeks, and in the two later years only about one week.
THE DATE OF MOULT IN THE JUVENILES IN RELATION TO AGE The date that the post-juvenile moult begins presumably depends mainly on some environmental factor, perhaps daylength, but also on the age of the bird, which might have left the nest anytime between the end of May and the beginning of October. It seems however that juveniles moult only after attaining a certain weight (see later). In general, juveniles which left the nest earliest in the season moulted first. In the summer of 1963, six birds ringed as nestlings were subsequently recovered in moult: four which left the nest in early June began to moult at the end of July, and two which left the nest in early July, began to moult about mid-August. In all years there were some juveniles examined after mid-August which must have started their moult within a fortnight of leaving the nest, since they still had under-developed rectrices and under wing-coverts, and some of which were also seen to be fed by their parents. None of these juveniles, however, had passed beyond stage 1 of moult, indicating only slight overlap of the growth of juvenile feathers with the start of post-juvenile moult. In birds which left the nest earlier in the season, several weeks elapsed between these two processes. A similar correlation between the age of the bird and the date of moult was recorded in the House Finch by Michener & Michener (1940), the Boat-tailed Grackle by Selander (1958), and in the Blackbird Tuvdus merulu by Snow (1958). As in the Bullfinch, the post-juvenile moult in the Andean Sparrow may begin before the rectrices are full- grown (Miller 1961).
THE MOULT AS A PERIOD OF PHYSIOLOGICAL STRAIN
THE RATE OF FEATHER SYNTHESIS DURING MOULT Although the moult is regarded as a period of physiological strain, the rate of feather synthesis and the food requirements during moult have not yet been studied for any species; but presumably, additional food is required then not only for the production of new feathers but also to compensate for extra heat-loss. The latter might result both from a reduction in insulation by the plumage and from the flow of blood through the growing feathers. Table 11 shows the weight of the feathers comprising different tracts, and the mean rate at which they were synthesised during moult. Although the feathers of the dorsal tract weigh twice as much as the primaries, and those of the ventral tract more than three times as much, their moults took less time, sometimes as much as a third. However, there are many more papiilae producing the feathers of the dorsal and ventral tracts than the flight feathers. It will be recalled that variation in the rate of moult in the remiges was achieved, not by changing the growth-rates of individual feathers, but by changing the number of feathers growing concurrently, a number limited, presumably, by the ill-effect their loss has on flight. It seemed plausible, however, that the prolonged moult in the remiges is due, not only to the need to maintain efficient flight, but to a particularly high energy demand. It can be seen from Table 11, however, that in the Bullfinch the remiges comprise only about a seventh of the weight of the entire plumage, so presumably the need for efficient flight is by far the most important factor causing their protracted moult. Thus, it is not surprising that the duration of moult in the primaries (usually 72-82 days) varied much less than that in the body feathers (usually 50-80 days of intense moult). 60 I. NEWTON : THE MOULT OF THE BULLFINCH PYRRHULA PYRREULA IBIS108
TABLE11. The mean wetght and moult periods of dayerent groups of feathers in the Bulljinch Pyrrhula pyrrhula. Primary score Mean length Total weight Mean weight at average of moult (by of feathers of feathers start and end primary score) replaced replaced per
of moult (mgm.)*,- unit increase in primary score (mgd Primaries 1-90 90 220 2.44 Secondaries 46-90 45 96 2.13 Tertials 27-69 43 26 0-58 Upper primary-coverts 1-90 90 20 0.22 Upper greater-coverts 13-32 20 32 1.50 Upper lesser- and median-coverts 8-48 41 38 0.93 Alula 65-78 14 7 0.50 Underwing-coverts 37-76 40 54 1.35 Shoulder tract 18-68 51 108 2.12 Ventral tract 13-85 73 736 10.08 Dorsal tract 13-80 68 490 7.21 Thigh tract 30-86 57 80 1*40 Legs 26-68 43 20 0.47 Head 19-70 52 86 1.65 Rectrices 27-79 53 136 2-59 Upper tail-coverts 34-80 47 28 0.59 Lower tail-coverts 30-77 48 24 0.50 * The figures are the averages of five birds; the range of values obtained was never more than 8% of the mean.
During moult, an average of 2.2 gm. of new feathers were produced by adult Bull- finches, and about 1.7 gm. by juveniles; this is equivalent to lo;/, and 8% respectively of their live-weights at the start of moult (see later), and 40% and 33% respectively of their dry-weights (Newton & Evans, in prep.). Since adult moult lasts 10-11 weeks and post-juvenile moult 7-9 weeks, the mean rate of feather synthesis by weight is about the same in both. Fig. 7 shows the rate at which new feathers were synthesised during adult and post- juvenile moults; this was calculated for the primaries, secondaries and rectrices of adults from Figs. 3, 1 and 2 respectively; in all other tracts (in the absence of other information) it was assumed to be constant throughout (see Table 11). Given this assumption, the production of feathers was slowest at the start and end of moult, but for most of the time averaged nearly 40mgm. per day (equivalent to 06% of the dry-weight of the bird) in both adults and juveniles. There was, however, considerable variation and in many birds feather production must, at times, have far exceeded this.
METABOLISM AND FOOD-INTAKE DURING MOULT In the domestic fowl the metabolic rate increases during moult by 45% (Perek & Sulman 1945), and in the Chaffinch by up to 25% (Koch & de Bont 1944). Wallgren (1954) found that in the Ortolan Bunting Emberiza horlulana the metabolic rate in adults adults increased by 26% during moult, and in juveniles, which replaced fewer feathers, by 10%; in the Yellowhammer Emberiza citrinella he showed an increase of 147; in adults, but none was detected in the juveniles. The higher standard metabolic rate in these species and the moult may have a common cause in increased thyroid activity (see King & Farner 1961). Little information is available on the relationship between the increased metabolism and food-intake during moult, but King & Farner (1961) have tentatively estimated that the House Sparrow needs 7.6% more food per day for feather synthesis, apart from any required for extra heat production. The feathers of birds consist almost entirely of 1966 I. NEWTON : THE MOULT OF THE BULLFINCH PYnmui,AiPYRRHIIIA 61
ADULT MOULT POST -JUVENILE MOULT Weight of feathers replaced (gms.) 1 ,2,3,LI 5, 6,7,8,9,10,11Time(weeks) 2.2 Weight feathers 2 .o of replaced (gms.)
1.7
1 -5 1 A5
1 .o 1.0
0.5 0.5
0 0
Stage of moult Stage of moult FIGURE7. The rate of feather synthesis during adult and post-juvenile moults in the Bullfinch Pyrrhula pyrrhula. (For explanation of moult stages see text.)
protein, and their constituent amino acids may be needed in proportions very different from those in which they are present in the proteins of natural foods. It is possible therefore that a larger weight of protein is required during moult than might be expected simply to supply enough of the less common amino acids. No relevant information is available for Bullfinches but it seems probable that extra food is required throughout the moult, especially since food-shortage on only one or two days is liable to weaken the structure of growing feathers.
ACTIVITY DURING MOULT Apart from eating more during moult, birds might conserve energy by resting for part of each day, especially if food is easy to obtain at this time. Captive Bullfinches, which normally sing or fly around for most of the day, when in moult were active only while feeding; the rest of the day was spent mainly in sitting or sleeping in a shaded position with feathers fluffed. They also bathed less and preened more than at other times.
TABLE12. The number of Bullfinches Pyrrhula pyrrhula caught at different stages of moult.
ADULTS Stage of moult 123456789 Number caught 43 27 22 14 9 22 9 7 29 Percentage of total caught 24 15 12 7 5 12 5 4 16
JUVENILES Stage of moult 1 2 3 4 5 6* Number caught 140 48 66 43 49 72 Percentage of total caught 31 11 16 10 12 18 Note. The table is based only on the data for 1963 and 1964 when trapping was carried out with equal intensity throughout the moult period. For explanation of stages see text. * Birds which had just completed the moult, as indicated by the presence of sheaths on sonie of the new feathers. 62 I. NEWTON : THE MOULT OF THE BULLFINCH PFRRHULA l’J*RRHL;Ld IBISio8
Table 12 shows the total number of adult and juvenile Bullfinches caught at different stages of moult, irrespective of the dates their moult occurred. Although each stage of moult took about the same time to complete, it may be seen that more birds were caught in the first and last stages of moult than in the middle stages when more feathers were being replaced. The difference is more striking in the adults than in the juveniles. This is consistent with a reduction of activity for most of the moult. Moreover, during any given period, birds in the first and last stages of moult could be caught at most hours of the day, but birds in heavy moult mainly in the early mornings and in the evenings. These data, however, indicate an average trend and refer mainly to birds moulting in the most favourable part of the season. Some individuals, especially those moulting late, may have to become even more active than normal in order to find enough food (see later). It was noticed also that relatively more moulting adults than juveniles were caught in nets placed in thick cover compared with those placed in more open situations, such as at patches of Meadowsweet. Such skulking behaviour, which was not apparent among moulting juveniles, would assist adults in heavy wing-moult to avoid avian predators. A similar difference in behaviour during moult between adult and juvenile Chaffinches was noted by Marler (1956).
WEIGHTS DURING THE MOULT PERIOD One might have thought that any strain imposed by the moult would either be compensated for by increased feeding or result in a loss of weight and even in death. Bullfinches in poor condition and markedly below average weight were not often caught, so the weight changes described below are typical mainly of healthy birds. Table 13 summarises the weights of adults during the moult period, and Table 14 those of juveniles, Seasonal changes in the weights of Bullfinches have been described in detail elsewhere (Newton, in press). Although both sexes lost weight while feeding their last young (Newton 1965), throughout the nestling period the females remained 1-2 gm. heavier than the males. But the weights of the males rose again with the onset of moult, suggesting that the first stage of moult (when only 1-2 primaries are growing in each wing) was less strenuous for them than parental care. The weights of the females however remained fairly constant from the end of the nestling period through to stage 2 of moult. By this time the weights of the two sexes were about equal; the weights of both then increased from stage 2 to stage 4, and thereafter remained about the same until the end of moult, when there might have been a slight further increase. The weights of the juveniles also increased during moult by some 2-3 gm. on average. This increase might have been due, in both adults and juveniles, partly to a deposition
TABLE13. The mean weights (with standard deviations) of adult Bulljinches Pyrrhula pyrrhula according to stage of moult (all years). August September October Stage Male Female Moult no: started (33) 21.7&1-5 (40) 23.2k1.6 1 (22) 22-3 2 1.2 (11) 23.1 f 1.1 2-3 (24) 22-921.2 (15) 22.821.0 (4) 23.0 4-5 (8) 23.7 (13) 23.8+ 1.0 6-7 (17) 23.7k1.0 (4) 23-8 8-9 (15) 23-82 1.3 Moult just finished* (18) 24.22 1.2 Notes. The figures in parentheses show the numbers weighed. Standard deviations are given only when more than ten birds in each class were weighed. * As indicated by the presence of sheaths on some of the new feathers. I966 I. NEWTON : THE MOULT OF THE BULLFINCH PYRRHULA I’YRRIfUL4 63
TABLE14. The mean weights (with standard deviations) of juvenile Bulenches Pyrrhula pyrrhula according to stage of moult. 1962 1963 September October August September October Stage Moult not started (4) 21.5 (3) 20.7 (26) 21.5 k 1-0 (3) 21.9 1 (5) 22-9 (13) 22.4k1.2 (30) 21-8k1.1 (18) 21*3+1.2 2 (2) 22.8 (8) 22.5 (9) 22.5 (9) 22.6 3 (3) 22.9 (7) 22-8 (8) 22.8 (5) 23.3 4 (6) 23.8 (11) 24-3 1.2 (6) 23.8 (8) 24.0 (5) 23.7 5 (1) 24.2 (10) 24.3 (1) 24.1 6* (5) 24.8 (22) 24.0 &- 0.9 (11) 23.9$11 .3 1964 August September October November Stage Moult not started (30) 21.4&1.0 (7) 21.1 1 (38) 21.8,tl.O (30) 21-611.1 L,. (4) 22.1 (8) 22.5 (2) 21.2 3 (10) 22.8 (16) 22.611.2 (9) 22.4 4 (3) 24.7 (8) 23.5 5 (8) 23.8 (15) 24.3&1-1 (3) 22.8 6* (14) 23.4,tl.O (24) 23.811.0 Notes. The figures in parentheses show the numbers weighed. Standard deviations are given only when more than ten birds in each class were weighed. * Moult just finished as indicated by the presence of sheaths on some of the new feathers. of fat, but in late-fledged juveniles, partly to general body growth as well. In both adults and juveniles the same range of weight changes occurred in different years and at different dates in any one year (Tables 13 and 14), as well as in particular individuals caught more than once during moult. There is therefore no evidence from these mean weight changes that the moult imposes a physiological strain on Bullfinches; but as mentioned above, the figures are almost certainly based on the weights of a sample of healthy birds. If any succumbed, the chances that they would be caught just beforehand are very small.
THE FOOD THROUGH THE MOULT PERIOD Table 15 shows the food of Bullfinches in Wytham Woods from August to November 1962-64; the data were obtained by repeated field observations (see Hartley 1953), checked by an occasional analysis of gut-contents (Newton, in prep.). In all years throughout the moult period the seeds of birch Betula spp., Meadowsweet, the bramble Rubus fruticosus and various docks predominated in the diet of Wytham Bullfinches; but the seeds of the nettle Urtica dioica were also important in the two later years. The seeds of various grasses were also frequently taken in August, and those of common privet Ligustrum vulgare in September.
TABLE15. The percentage composition of the food of the BullJinch Pyrrhula pyrrhula in Wytham Woods, Oxford, from August to November in three successive years. 1962 1963 1964 Aug. Sep. Oct. Nov. Aug. Sep. Oct. Nov. Aug. Sep. Oct. Nov. No. of obs. 101 162 215 164 102 134 152 131 117 153 201 102 Birch 4513 36 29 34 20 16 28 14 25 20 15 7 Grasses 1 0 0 16 2 0 0 14 2 0 0 Docks 6 174 4 6 15 17 4 16 14 10 12 16 Bramble 9 17 21 12 18 20 25 10 11 14 18 Meadowsweet 17 20 20 3 19 24 26 2 24 26 13 2 Wood avens 4 3 1 0 4 5 1 0 3 5 1 0 Nettle 2 3 8 7 3 8 17 39 3 12 40 55 Elder 1 1 150 0 1 1 0 0 1 2 1 0 Privet 0 14 6 0 4 0 0 0 8 2 0 Ash 0 0 1 20 0 0 0 0 0 0 0 1 Othem 3 1 5 5 10 5 4 4 6 4 2 1 64 I. NEWTON : THE MOULT OF THE BULLFINCH P'YRHHIJLAPYRRHULA IBIS108
In Wytham Woods, the seeds of various herbaceous plants were abundant and continually replenished from May to September each year. It is unlikely therefore that food for Bullfinches was short at this time, though no measurements were made. From October however, there was no fresh growth, and Bullfinches fed mainly on seeds which were to last them into the winter, namely those of bramble, docks, nettle, birch, and privet. The seeds of ash were also important, but these were not taken in any quantity until they had ripened after the end of the moult period; they were plentiful in 1962 and 1964. The crops of birch and privet seeds were largest in 1962 and smallest in 1963. These seeds formed a higher proportion of the diets of Bullfinches in 1962, when they were most plentiful, than in the following two years (Table 15).
SURVIVAL IN RELATION TO DATE OF MOULT Estimates were obtained of the number of Bullfinches in Marley wood (by a capture- recapture technique), and of the seeds available to them, each month from October to March over three successive winters. These data, together with the sampling methods, are given elsewhere (Newton 1964, 1964 a, and in prep.), but certain points are relevant here. In all years, almost all the available seeds of birch, privet, and nettle had gone by mid-December, and in the autumn of 1963, when the birch and privet crops were small, a large proportion of the dock and bramble seeds had also disappeared by then. Also, in all years the number of Bullfinches in Marley fell from October to December, but the decline was greatest in 1963 when seeds were scarcer. In all years, a greater percentage of juveniles than of adults disappeared; this was reflected both in the recapture data and in the proportion of juveniles in the trapped samples, which fell markedly over this period. It is thought that most of the birds which disappeared each autumn had died, since the changes in numbers recorded in Marley were apparently typical of those occurring over a much wider area (Newton 1965 a). Further, only a few birds (all juveniles) were subsequently recovered elsewhere; none of the adults ringed in Wytham Woods have yet been recovered more than a few hundred yards away. The period from early October to mid-December, when some birds were still in moult, was therefore a period when food became progressively scarcer, when tempera- tures dropped and days shortened, and when many birds disappeared (though the losses varied in different years). It seems likely that at least in some years, birds still moulting in late autumn would have a smaller chance of survival than those which had finished their moult, a disadvantage to be weighed against the fact that adults late in moulting can raise an extra brood of young (Newton 1964). In each year there was a period when birds at widely different stages of moult occurred in the population together; but at any one time, birds at late stages of moult were always heavier than those at earlier stages (Tables 13 and 14) and may therefore have stood the best chance of survival. To avoid this possible bias in comparing the survival (=recovery) rates of birds beginning to moult at different times in the season, I have used only those caught in the first stages of moult, before any severe strain is likely to have been incurred. In addition, only birds caught initially in Marley have been used, since trapping was restricted to this area during the rest of the year. The subsequent recoveries refer to birds caught after the end of the moult season, i.e. after November.
(1) Adults It is unfortunate that in 1962 netting was not started until mid-September since in this year there was prolonged breeding by some Bullfinches and a considerable spread in the date on which moult started. In this year, the difference in survival between adults moulting early and late might well have been less than expected, since birds presumably continue breeding and moult late only in exceptionally favourable years when the seed- crops, produced in late summer, are large enough to remain abundant into the autumn 1961 I. NEWTON : THE MOULT OF THE BULLFINCH IJYRRHULAPYRRIIC'LA 65
(see later). Thus in 1962 almost all the late-moulting adults caught were recovered after the end of the moult season, which showed that they had not suffered appreciable mortality as a result of their late moult. In 1963, data for analysing survival in the adults are insufficient. In 1964, however, five out of eight birds that started to moult in the first half of the season were subsequently recovered, compared with one out of five birds that started to moult in the latter half. This suggests that in 1964 the adults moulting later survived less well than those moulting earlier, though more data are required to substantiate this. The alternative explanation, that more of the late-moulting adults had left the wood, is unlikely, since as mentioned above, there are no records of Wytham adults moving very far.
(2) Juveniles Table 16 shows the recoveries of juveniles beginning to moult in different months 1962-64. Only birds caught in the first stage of moult are included. Further, to exclude birds which might still have been dependent on their parents when they started moult, all those with under-developed rectrices when caught were omitted from analysis. Many juveniles were later recovered in Marley, but three were reported from outside the wood (within two miles) by members of the public and are listed separately. It may be seen that in all years more of the juveniles moulting early were recovered after completing their moult than of those moulting late. Though the data are few, they suggest that late moulting carries a lower survival (=recovery) rate for juveniles than early moulting. The alternative explanation, that more of the late moulting birds had left the area, is again unlikely since all those recovered outside Marley were birds which had moulted early. TABLE16. Survival of juvenile BultJSnclaesPyrrhula pyrrhula in relation to time of moult. Start Number Recaught after November moult examined in Marley elsewhere Total Percentage 1962 September 9 4 1 5 56 October 14 4 0 4 28 1963 July 4 3 1 4 (1 5 5 August 18 7 1 8 September 6 2 0 2 33
1964-. .. July 5 1 August 47 18 September 39 7 It is not surprising that the early moulting birds survived better, since, in addition to moulting at the most favourable time, they were by then more experienced at finding food, for they had already been out of the nest 4-8 weeks compared with only 2-3 wecks for the late-moultingjuveniles. It is of interest also that in 1962, when the late-moulting adults survived well, the late-moulting juveniles apparently suffered heavy losses. This is again presumably because these juveniles were younger, less experienced at finding food and began moult 2-3 weeks later than their parents. Finally, it is worth stressing that in all years most birds moulted while food was still plentiful; only a few (mainly juveniles) were still in heavy moult when food became scarce. These observations therefore support the view that the moult is timed to occur while food is still plentiful. Since most of the seeds eaten by Bullfinches during moult are already available and are eaten to some extent earlier in the summer, the birds have presumably evolved the ability, as in 1962, to prolong breeding and simultaneously postpone moult in those years when food is exceptionally plentiful. But ewn in 1962, a proportion of birds finished breeding and moulted at the normal date. It is possible E VOL. 108 66 I. NEWTON : THE MOULT OF THE BULLFINCH PYRRHC'LA PYRRHC'LA IBIS108 therefore that selection to avoid a late moult preserves some birds which have a hereditary tendency to moult early, irrespective of conditions.
DISCUSSION In normal years in England, a period beginning at the end of July, during which food is abundant, is probably the most favourable time for moult in the Bullfinch (the food-supply and the timing of the breeding season being the " ultimate " factors to which the timing of moult is adapted). Juveniles which left the nest after about mid- August began to moult within a fortnight, suggesting that the advantage of moulting then outweighed that of waiting until they were well grown and more experienced at finding food. Doubtless, therefore, juveniles which left the nest at the end of May could have begun moult before the end of July if it had been advantageous for them. Further- more, even the non-breeding males did not begin moult until towards the end of July. Presumably there is some external regulator, perhaps daylength, which acts as the main proximate factor initiating the moult at this date. But the effect of daylength is modified in the adults by the date at which breeding activity ends, and in most of the juveniles by their age. The abundance of food and warm weather may not be the only factors making the end of July the best time for all Bullfinches, adults and juveniles, to begin moult. Many of the adults examined just before moulting late in 1962 had very worn plumage (see above), and these were birds which cannot have moulted for more than a year. Even longer periods might elapse between successive moults unless selection kept the timing of moult fairly constant from year to year; the plumage of any bird, adult or juvenile, which moulted early in the summer of one year (before the end of July) might be so worn by the end of a long breeding season the following year as to impair flight and insulation. Natural selection will presumably favour those individuals which raise most broods during a season. If moult occurs at a time of year when breeding is also possible (but this has not been proved for the Bullfinch), selection will tend to compress the moult to within as short a time as possible in the latest part of the period during which food is abundant, a period which varies in duration from year to year. If food were continuously plentiful, the latest date at which a moult could be safely accomplished in a resident species, such as the Bullfinch, would probably be set by the colder and shorter days of winter when insulation by the plumage is most needed. In all three study years, most Bullfinches finished moult just before food became scarce, even though this occurred later in one year than in the other two. Although in some species moult is markedly slower when food is scarce (Ashmole 1962), even with abundant food there will obviously be a limit to the rate of moult; it must not proceed so fast as to hinder flight or serioudy reduce insulation.
ACKNOWLEDGMENTS My sincere thanks are due to Dr. P. R. Evans and Mr. R. E. Moreau for criticising the manuscript and for their many excellent suggestions. P. R. Evans gave particular help with the statistical analyses. I am also indebted to Dr. D. Lack and Dr. C. M. Perrins for useful comments on the manuscript, and have benefited greatly from discussion with Professor A. J. Cain and Dr. D. W. Snow. This work was carried out while I was holding grants first from the Department of Scientific and Industrial Research, and then from the Agricultural Research Council. SUMMARY The distribution of feather tracts and their sequence of moult in the Bullfinch is described. The adult post-nuptial moult, which is complete, lasted 10-12 weeks, and the post-juvenile moult, which is partial, 7-9 weeks. Adult moult began with the shedding of the first (innermost) primary and ended with the replacement of the last. Variations in the rate of moult in the flight feathers were mainly achieved, not by changes in the growth rates of individual feathers, but in the number of feathers growing concurrently. The primaries were shed more slowly, and the onset of body moult delayed, in birds which were still feeding late young. 1966 I. NEWTON : THE MOULT OF THE BULLFINCH PYRRHL'LA I'l'RRRULA 67
In 1962, the onset of moult in the adults was spread over 11 weeks from thc end of July to the beginning of October, and in the two following years over the six weeks, from the end of July to the beginning of September. The onset of moult was delayed by late breeding, which itself occurred in response to a comparative abundance of food in late summer, markedly in 1962. In all years, the first juveniles to moult started at the end of July, and the last, three weeks after the latest adults. Juveniles moulting late in the season retained more juvenile feathers than those moulting earlier. During moult, adult and juvenile Bullfinches produce feathers equivalent to 40':: and 33% respectively of their dry weights. In both, for much of the moult, an average of nearly 40 mgm. of feather material-some 0.6"; of their dry-weight-is laid down each day. The remiges of the adult comprise only a seventh of the weight of the entire plumage, and it is suggested that their protracted moult results not so much from their energy requirements, as from the need to maintain efficient flight. Variation in the rate of moult in the remiges was much less pronounced than in the body feathers. Bullfinches were less active during moult than at other times of the year. The weights of both adults and juveniles increased during moult. The food during the moult period is described. In all years, most Bullfinches finished moulting just before food became scarce, even though this occurred at different times in different years. In one year, adults moulting latest in the season probably survived less well than those moulting earlier; the same was apparently true of the juveniles in all years. The timing of moult in the Bullfinch, and the factors initiating it, are discussed in relation to the breeding season and food- supply near Oxford. REFERENCES ASHMOLE,N. P. 1962. The Black Noddy Anous tenuirostvis on Ascension Island. Ibis 103b: 235-273. BLANCHARD,B. D. 1941. The White-crowned Sparrows (Zonotrichia leucophrys) of the Pacific seaboard: environment and annual cycle. Univ. Calif. Publ. Zool. 46: 1-178. DIXON,K. L. 1962. Notes on the molt schedule of the Plain Titmouse. Condor 64: 134-139. DWIGHT,K. L. 1900. The sequence of plumages and moults of the passerine birds of New York. Ann. hTewYork Acad. Sci. 13: 73-360. EATON,S. W. 1957. A life history of Seiuvus noveboracensis. Sci. Studies, St. Bonaventure Univ. 19: 7-36. EVANS,P. R. In press. Autumn movements, moult and measurements of the Lesser Redpoll Carduelis jlammea cabaret. Ibis 108. GIBB,J. A. 1954. The feeding ecology of tits with notes on Treecreeper and Goldcrest. Ibis 96: 513-543. JOHNSTON,D. W. 1961. Timing of annual molt in the Glaucous Gulls of northern Alaska. Condor 63 : 474-478. KING,J. R. & FARNER,D. S. 1961. Energy metabolism, thermoregulation, and body temperature. In ' Biology and Comparative Physiology of Birds '. (Ed. J. A. Marshall.) Vol. 2: 215-288. London & New York: Academic Press. KOCH,H. J. & DE BONT,A. F. 1944. Influence de la mue sur l'intensite de mktabolism chez le Pinson Fringilla coelebs coelebs L. Ann. SOC.Zool. Belg. 75: 81-86. MARLER,P. 1956. Behaviour of the Chaffinch Fringilla coelebs. Behav., suppl. 5: 147-150. MICHENER,H. & MICHENER,J. R. 1940. The molt of House Finches of the Pasadena region, California. Condor 42: 140-153. MILLER,A. H. 1928. The molts of the Loggerhead Shrike Lanius ludovicianus Linnaeus. Univ. Calif. Publ. Zool. 30: 393417. MILLER,A. H. 1933. Postjuvenal molt and the appearance of sexual characters of plumage in Phainopepla nitens. Univ. Calif. Publ. Zool. 38. MILLER,A. H. 1961. Molt cycles in equatorial Andean Sparrows. Condor 63: 143-161. NEWTON,I. 1964. The ecology and moult of the Bullfinch. D. Phil. thesis, Oxford University. NEWTON,I. 1964 a. Bud-eating by Bullfinches in relation to the natural food-supply. J. Appl. Ecol. 1: 265-279. NEWTON,I. In press. Seasonal fluctuations in the weights of Bullfinches. Brit. Birds. PEREK,M. & SULMAN,F. 1945. The basal metabolic rate in molting and laying hens. Endo- cinology 36: 240-243. PITELKA,F. A. 1958. Timing of molt in the Steller Jays of the Queen Charlotte Islands, British Columbia. Condor 60: 3849. RIDDLE,0. 1908. The genesis of fault bars in feathers and the cause of alternation of light and dark fundamental bars. Biol. Bull. Woods Hole 14: 328-370. SELANDER,R. K. 1958. Age determination and molt in the Boat-tailed Grackle. Condor 60: 355-376. SELANDER,R. K. & GILLER,D. R. 1960. First-year plumages of the Brown-headed Cowbird and Red-winged Blackbird. Condor 62: 202-214. SNOW,D. W. 1958. A Study of Blackbirds. London: Allen & Unwin Ltd. SOUTHERN,H. N. 1954. Tawny Owls and their prey. Ibis 96: 384-410. WALLGREN,H. 1954. Energy metabolism of two species of the genus Emberiza as correlated with distribution and migration. Acta Zool. Fenn. 84: 1-110. WRIGHT,P. L. & WRIGHT,M. H. 1944. The reproductive cycle of the Red-winged Blackbird. Condor 46: 46-59. Dr. I. Newton, Edward Grey Institute, Botanic Garden, Oxjord.