PLEASE SCROLL DOWN FOR ARTICLE

Transcription

1 This article was downloaded by: On: 16 September 2009 Access details: Access Details: Free Access Publisher Taylor & Francis Informa Ltd Registered in England and Wales Registered Number: Registered office: Mortimer House, Mortimer Street, London W1T 3JH, UK Bird Study Publication details, including instructions for authors and subscription information: The breeding biology of the Chaffinch I. Newton a a Edward Grey Institute of Field Ornithology, Oxford Online Publication Date: 01 March 1964 To cite this Article Newton, I.(1964)'The breeding biology of the Chaffinch',Bird Study,11:1,47 68 To link to this Article: DOI: / URL: PLEASE SCROLL DOWN FOR ARTICLE Full terms and conditions of use: This article may be used for research, teaching and private study purposes. Any substantial or systematic reproduction, re-distribution, re-selling, loan or sub-licensing, systematic supply or distribution in any form to anyone is expressly forbidden. The publisher does not give any warranty express or implied or make any representation that the contents will be complete or accurate or up to date. The accuracy of any instructions, formulae and drug doses should be independently verified with primary sources. The publisher shall not be liable for any loss, actions, claims, proceedings, demand or costs or damages whatsoever or howsoever caused arising directly or indirectly in connection with or arising out of the use of this material.

2 The breeding biology of the Chaffinch By I. Newton Edward Grey Institute of Field Ornithology, Oxford Received 29 April 1963 INTRODUCTION THIS PAPER SUMMARISES data from over 3,500 nest record cards for the Chaffinch (Fringilla coelebs), from the start of the nest record scheme in 1934 up to, and including Data on over 30o nests has also been extracted from the diaries of the late Arthur Whitaker, which have been deposited in the Edward Grey Institute, Oxford. Whitaker collected his data during the first half of the present century, mainly from an area on the Derby-Yorkshire border. The chief types of bias inherent in the nest record scheme are described separately under each section, but it is worth stressing at the outset that the results obtained from such an analysis are greatly influenced by the way in which the data are collected and analysed, and especially by the way in which allowance is made for the different types of bias. Because of this, absolute values for the figures obtained are of far less significance than their comparative values. Clearly nesting success could vary greatly according to local conditions, and habitat. In this respect, it is worth noting that over half of the Chaffinch nests recorded were in gardens. Unfortunately there are insufficient cards available for woodland to make a comparative analysis of the two habitats worth while, but it was apparent that there were slight differences in the breeding season, and success rate, for birds nesting in different habitats. The present work owes much to previous analyses, in particular to that of Myres and Snow (1955), on whose papers the present one is largely modelled. BREEDING SEASON The date on which the first egg of a clutch was laid has been taken to indicate the onset of breeding. Taking this, rather than the date of clutch completion as the critical point in the timing of breeding, more cards could be used. This is because many nests were not visited after the first or second egg. Where this date has not been recorded on the cards, it has been estimated as follows : t) For clutches found during laying, the date of first egg could be inferred exactly, or almost exactly by backdating, eggs normally being laid on consecutive days. 2) In records in which the date of hatching was known, the date of first egg could be found by subtracting thirteen days plus the clutch size D47

3 PERCENT OF TOTAL CLUTCHES STARTED 20 BIRD STUDY minus one. (Thirteen days is the average incubation period.) This estimate is accurate to within two days. 3) If the fledging date was given, a less accurate estimate was obtained by subtracting twenty-seven days plus the clutch size minus one. (Twentyseven days is the sum of average incubation and fledging periods.) Similarly the date of first egg could be estimated when the age of the young was given on the card. Only cards recording laying, hatching or fledging dates, or the age of the young, have been used in this analysis. Data on the start of breeding were available from a total of 2,510 nests. For most purposes, the dates of first egg have been grouped into five day periods. The use of a ten or a seven day grouping can obscure seasonal and local variations, and may even be misleading. (Myers, 1955.) With many species, a higher proportion of nests of earlier breeders is probably found than of birds nesting later, because at the beginning of the season, enthusiasm for nest finding is greatest, and cover least dense. However, since the Chaffinch has a relatively early and short breeding season, the data should not be seriously biased in this way. Moreover, records from the diaries of Whitaker, an ornithologist who searched more diligently throughout the season than most observers, give an almost identical pattern of breeding to that indicated by the nest record cards A M F IGURE 1. The breeding season of the Chaffinch, as shown by the percentage of total clutches started in each five-day period, throughout the season. Data are incorporated from the whole of Great Britain, from 1934 to 1960 inclusive. 48

4 BREEDING BIOLOGY OF CHAFFINCH 1q.._ ab oob co,,-00 no,n moo 0 O /7^0.L 0, N O, MN m oo b ^ n D\b yn N N N m N w T x ^ F s^ $_b O O O O O - n a 0 ro o :'z-oz N.. O O O O et n a, ^ w 6r-fr o..onooo. 000c, ^ ó, B O H 6r-or ^-. N.. 0 v, m.... O 0 0 Hh N 0 a -S N^ n N.. Nen v,.... O.. M N 06 B ~ O fr-r v, N b m 0 d m, m N 0.00 Ǹ O T 0 2 G o -9z x N,n ,0 N.. 0 nb +^ B. ó ^ F V z 0 6z-SZ Op N "z-rz oo\ h O oo O\ O -I- N O ^ N o Moo. M. V m.tn G á y w z v w v 0z 9T N O\ ^ O oo h hb N b N moo ^ i, C.. U C F^ Q ai, w S r-r l b o0 07.N. ^ \Ó.n N N M.N. N N m N w N > l' 6 ty w w m 0 N b ^ N,O N ^, 00 n.. m D\ aor--9 :,:-1 N en en N N en.... v, 0 xi 0 M U d) w w 0S-T coo -a.^ G\ ^ M O, -1-F..n en ^.. a a0i - en N en en.4-m.. Y,/, N N m.o Z.. v^ Q O -.9Z en N '^d N N N M N^.N. N m m.. 07 en 0 HY, w' U <Sz-TZ "ȧ, b n 04 H M oh H N N oo 0, N co O I. N ^ h ^ s1 Á. O w ^,M^ oz-9, O mv,.. neyd. N ^ O N O, m N N M e, W B 5.I U iy HSI-rr "IN N N...n en 0 N en N m.. V, b N O b y.7.0 U OI-9 7 }, N'.. N.+.. co.. N O O-.I. H O S-r ,10000,O ai.. w ^^ r -LZ O O O O O O O O N O N 5 ^ 9z-zz w ON. a 0.1 ^ O.. N en noo ON O m n v, v, v, v, v, v, v, v,b u F a a"d\ aá 0\ 0, ONO.O,^ 0 I. 49. h

5 BIRD STUDY THE PATTERN OF BREEDING In most years, the first few clutches are completed before mid-april, and the last during June; there is a pronounced peak in egg laying which on average, occurs during the first few days of May (Fig. 1). In some years, for instance in 1954 and 1958, there was evidence also of a smaller, second peak, which occurred two to three weeks after the first peak (Table I). Extremely early clutches were reported in the spring of 1957, the earliest being started on 17 March. The latest clutches reported were one in 1960, begun on 12 July, and one in 1949, started on 3 August. There is little indication from the nest record cards of a regular second brood, though a few alleged genuine second broods were reported for Chaffinches nesting in gardens; in such cases, the second clutches were often laid in the same nests as the first. Repeat clutches are commonly laid, and Whitaker records a third repeat, after failure of the first three clutches. In any one year or locality, the main peak was usually spread over a shorter period than Fig. I would suggest, and was sometimes confined to within a period of fifteen days or less. The most extreme case was in 1958, although relatively few clutches were reported. In this year, the first egg in about 4o per cent of the 129 recorded nests occurred within one five day period (see Table I). Thus, the broad peak of Fig. 1 really represents a series of narrower annual and regional peaks. In addition, the pattern of breeding was not identical in different years, but varied considerably according to the prevailing temperatures. This will be dealt with in more detail later. REGIONAL DIFFERENCES IN THE TIMING OF BREEDING In Fig. 2, the laying season of the Chaffinch has been plotted separately for different regions. The country has been divided arbitrarily into south-east, south-west, northern England, and Scotland. The southern boundaries of Cheshire, Derbyshire, Nottinghamshire, and Lincolnshire have been taken as the line dividing south from north, and the eastern boundaries of Warwickshire, Gloucestershire, Wiltshire, and Dorset as the line dividing south-east from south-west. These follow regions used by Summers-Smith (1952). Chaffinches breeding in the south-west are on average, a few days ahead of those nesting in the south-east and north England, and a fortnight ahead of birds breeding in Scotland. Although the peak laying period for the south-east and north corresponds, many more clutches are begun in the preceding ten day period in the south-east than in the north. Such a difference between the average laying dates in different regions is presumably due to the milder and earlier spring in the south and west of the country. 50

6 BREEDING BIOLOGY OF CHAFFINCH PERCENT OF TOTAL CLUTCHES STARTED SW n A ''^^^^ ' A N SCOTLAND 10-5 FIGURE 2. Regional differences in the timing of breeding, as shown by the percentage of total clutches started in each five-day period, throughout the season. Data are presented separately for the south-west, south-east, northern England, and Scotland, and covers the period 1934 to r96o inclusive. 5 1

7 BIRD STUDY ANNUAL DIFFERENCES IN THE TIMING OF BREEDING There are sufficient cards for the Chaffinch to demonstrate annual differences in laying time. In Table II, the date of the earliest clutches, and the peak laying dates are listed for the south of England, from 1950 to 196o inclusive. (The `earliest' clutches were arbitrarily taken to be the first five per cent of all the nests reported for each year.) TABLE II THE DATE OF THE FIRST FIVE PER CENT OF EARLY CLUTCHES, AND THE PEAK LAYING DATES, FOR THE SOUTH OF ENGLAND, FROM , INCLUSIVE Year Peak laying period Five per cent nests started by: May 20 April May 25 April o April 17 April May 20 April April is April 1955 i-5 May 20 April 1956 r-5 May 22 April April 5 April May zo April April Io April April 21 April The date by which five per cent of the clutches had been started varied by as much as a fortnight in different years. The peak period of laying is similarly variable from year to year, but usually occurs within the period 21 April to 5 May. The peak bears no constant time relation to the date of first clutches, but generally occurs within the following fortnight. Finally, in Fig. 3, the data for the two extreme years have been plotted side by side, 3957 being a very early year, and 1953, a rather late one. The 1957 season was exceptional, and an extremely early one for many species (see Lack and Goodacre, 1959). In the Chaffinch, both the earliest clutches and the peak date of laying occurred about a fortnight earlier than in an average year. THE INFLUENCE OF WEATHER ON BREEDING In temperate regions, change in daylength appears to be the only environmental factor which is constant from year to year, and which could therefore control gametogenesis, and the onset of breeding in birds. However, changes in daylength alone are insufficient to control the actual onset of egg laying, because it varies annually, according to the earliness or lateness of the spring. It is clear that in many bird species, the final control of the timing of breeding depends at least partly on air temperatures in the period prior to egg-laying. It is almost certain that each species is physiologically adjusted to respond to a definite temperature range, such that the young are produced at a period most propitious for their survival. 52

8 BREEDING BIOLOGY OF CHAFFINCH PERCENTAGE OF TOTAL CLUTCHES STARTED M A M FIG U RE 3. Differences in the timing of breeding in the south of England between the years 1953 and 1957, as shown by the percentage of total clutches started in each five-day period throughout the season. In the following section, some attempt has been made to correlate breeding activity in the Chaffinch with spring temperatures. The weather data are for the Oxford district, and have been supplied by the Geography School, University of Oxford. Nest records for the south-east of England only have been used, because as expected, they follow the Oxford weather data more precisely than do records for the country as a whole. In general, the method of analysis follows that developed by Myres (1955). Not every year's records have been analysed; data are presented only for the seven years for which most nest record cards are available. The 'full data are set out in Fig. 4, in which the number of nests, and mean air temperatures are plotted for each five day period. DETAILED DESCRIPTION OF THE DIFFERENT YEARS The first two clutches were started during the second five day period in April, when mean temperatures over the previous five days had averaged 45 F. However a subsequent drop in temperature below 45 F., at the end of April, again temporarily checked breeding activity. No correlation with temperature appeared after about the first week in May, after which temperature rose steadily A rapid rise in mean temperature from about 35 F. at the end of March to over 50 F. by mid-april resulted in a sharp increase in breeding activity. Again, egg-laying began when temperature increased above 45 F. It is difficult to be certain what effect the drop in temperature at the end of April had on breeding, but it may be relevant that the minimum temperature was well above, the minimum required to initiate breeding. 53

9 ^ f _1 r- U N a 2 w O O.> S\ \ r` w. = w ^^ U N m. ^^ n O ^ ó ^ VI c, = W r ^^ FIGURE 4. The correlation between the pattern of breeding and the prevailing temperatures, in different years. Both the mean temperatures and the number of clutches started, are plotted in five-day periods throughout the season. The broken line refers to temperature; the unbroken line to clutches. 54 D< U N N O

10 BREEDING BIOLOGY OF CHAFFINCH A rather slow rise in spring temperatures resulted in a steady rise in breeding activity, but a sudden cold snap at the end of April inhibited egg laying in many birds, and caused the peak for this year to be rather later than usual. Although a mean temperature of about 45 F. is sufficient to initiate breeding in April, no nests were reported in March 1953, when temperatures were well above 45 F The range of temperature variation in this year was comparatively slight, and appeared to have little or no influence on the pattern of breeding. Temperatures at the end of March were higher than average, and breeding started about five days earlier than usual As in and 1954, temperature exceeded 45 F. at the end of March, but no breeding occurred. In this year, the rather largefluctuations in temperature affected breeding almost to the end of May. The number of clutches started in each five day period was influenced by temperature over the previous five day period As in 1953, a slow rise in spring temperatures caused a very gradual increase in breeding activity, which was temporarily checked by a cold snap at the end of April In this year, there was an extremely early spring, with February temperatures consistently above 5o F., about 12 degrees higher than average for the month. The first nests reported were started in mid- March, and breeding activity then gradually increased despite a fall in temperature to about 46 F. in April. From the above account, the following tentative conclusions may be drawn : 1) The marked variations in the pattern of breeding in different years are correlated with the prevailing temperatures. The number of clutches begun in each period in April, and usually at least the first two periods in May, appears to be correlated with the temperatures over the previous five day period. In 1955, temperature influenced breeding almost to the end of the season. 2) With a steep rise in spring temperatures, there is a sharp increase in breeding activity, and a relatively narrow peak in egg-laying (see the data for 1952). With a slow rise in spring temperatures, the start of breeding in the population as a whole is more prolonged, and results in a less pronounced and broad peak in egg-laying (see the data for 1956 and 1953) 3) Most probably it is the absolute spring temperatures, rather than relative change in temperature to which the birds are responding, and there is much variation in the temperature required to initiate egglaying in individual birds. The threshold temperature in April that has to be reached before breeding will begin is about 45 F. This theshold temperature is higher for March (see data for 1953, 1954, and 1955), and lower for the end of April and May (see data for 1953 and 1956). In the unusually early spring of 1957, breeding was correspondingly earlier, and had started by mid-march. 55

11 BIRD STUDY 4) A sudden cold snap may inhibit breeding in many individuals, the most instructive example occurring in The adaptive significance in this response is not immediately clear since the weather at the time of egg-laying bears no relation to the weather at the time the young are in the nest. The most probable explanation seems to be that suggested by Dr. D. Lack (pers. comm.), that since low temperatures result in a delay in the flush of insect-food, delay of breeding in the Chaffinch is an adaptation acting to bring the nestling period into closer correlation with the food supply. ADAPTATION IN THE TIMING OF THE BREEDING SEASON In temperate regions, the breeding season of many birds is so timed that the maximum supply of appropriate food is available when the young are in the nest, and for a sufficient time after they have fledged. Chaffinches breed from the end of April until June, but there is a pronounced peak in egg-laying which usually occurs during the first few days of May. Most young Chaffinches are therefore in the nest during the latter half of May and early June, a period which corresponds with the peak availability of insect food, on which the young are reared (personal observations). Air temperatures greatly influence the pattern of breeding in different years, and by initiating or inhibiting breeding, serve to bring the nestling period more nearly to coincide with the peak abundance of insects. INCUBATION AND FLEDGING PERIODS The incubation and fledging periods were calculated, each for zoo clutches laid prior to 195o, and the results are set out on a percentage basis in Table III. TABLE. III-THE PERCENTAGE DISTRIBUTION OF THE INCUBATION AND FLEDGING PERIODS FOR THE CHAFFINCH Length in days 9 io Incubation period 1 % 4% 13% 21% 37% 14% 6% 4% Fledging period 6% 12% 23% 29% 17% 7% 2% 3% 1 % I) The incubation period varied from 9 to 16 days, with a mean period was 12.6 days (standard deviation 1.3). 2) The fledging period varied from II to 19 days, and the mean period was 13.9 days (standard deviation 1.6). CLUTCH SIZE The recorded clutch size was accepted for inclusion in this analysis only from two categories of cards : I) For nests which were visited daily over the end of the laying period and beginning of the incubation period. Here, clutch size could be 56

12 BREEDING BIOLOGY OF CHAFFINCH determined with some certainty, as the chance of an egg being lost before it was recorded is small. 2) For nests visited during incubation. (a) If two consecutive visits more than twenty-four hours apart showed no change in the number of eggs. (b) If the nest was visited once during incubation, but a sufficient number of days after recording an incomplete clutch, to be certain that the full clutch was present. It is possible that one or more eggs could have been lost before the observer visited the nest, this possibility increasing with the degree of incubation. However, Lack (1948) has shown in the case of the Robin (Erithacus rubecula), that the effect of this source of error is negligible. Although data on clutch size are open to slight error, they are not susceptible to the observational bias which influences an assessment of the breeeding season and nesting success. There is for instance, no reason why these sources of error should affect one year, season, or region more than any other. Also, so far as is known, clutch size does not affect the ease with which nests are found. Thus, on the basis of nest record card data, any annual, seasonal, or regional differences in clutch size are comparable. The mean clutch size for the whole country over seventeen years was 4.3 eggs (standard error 0.004). This mean, based on a total of 1,014 clutches, is probably a slightly low estimate, because of the effects of possible losses mentioned above. However, data on 368 clutches from Whitaker's diaries, give effectively the same mean clutch size of 4.4 eggs., TABLE IV-THE FREQUENCY DISTRIBUTION OF THE DIFFERENT CLUTCH SrZ$S, AS INDICATED BY THE NEST RECORD CARDS AND WHITAKER'S DIARIES Clutch size Number of.clutches Total Percentage ,182 From Table IV, which shows the frequency distribution of different clutch sizes, it will be seen that the majority of clutches consist of 4 or 5 eggs. There were about a quarter as many clutches of 3, and very few clutches of 2 and 6 eggs recorded. It is likely that many of the clutches of 2 are the result of one or more eggs being lost before discovery. Seven eggs in a nest were recorded only twice. SEASONAL VARIATION IN CLUTCH SIZE It is now well known that many species of birds show a seasonal change in mean clutch size, either a gradual decrease as the season progresses, or with a mid-season peak. Such changes have been correlated with the 57

13 BIRD STUDY amount of food available for the young, and are therefore adaptive (Lack, 1954). Table V shows the mean clutch size of the Chaffinch for each ten day period throughout the season, and covers data for seventeen years. In this species, clutch size remains constant until the end of May, after which it declines. There is no clearly defined mid-season peak of larger clutches, as occurs in certain other species (Snow, 1955). Table V also TABLE V THE PERCENTAGE COMPOSITION OF CLUTCH SIZES AND THE. MEAN CLUTCH SIZE IN TEN-DAY PERIODS THROUGHOUT THE BREEDING SEASON Clutch size Date No. of clutches 2 or or 6 Mean To April 5.. II Io% 55% 35% (4.4) April % 44% 53% 4.3 April % 45% 47% 4.4 April 26-May % 49% 43% 4.4 May % 45% 43% 4.4 May % 46% 45% 4.4 May 26-June % 54% 3o % 4.1 June % 53% 20 % 3.9 After June % 25% 17% (3.5) Note: Brackets indicate a lack of data and these means are of limited reliability. shows the relative frequency with which the different clutch sizes occur throughout the season. For convenience, clutches of 2 and 3, and of 5 and 6 eggs have been included together. The relative proportion of different clutch sizes remains about the same until the end of May. Then the proportion of 5 or 6 egg clutches decreases greatly, and that of 4- egg clutches slightly, with a corresponding increase in the proportion of 3- egg clutches. For most of April, and the first half of May, 5 eggs is the commonest clutch size. Clutches of 4 eggs predominate at the start of the season, and from mid-may until mid-june. Only at the very end of the season do 3- egg clutches predominate. ANNUAL VARIATIONS IN MEAN CLUTCH SIZE Mean clutch size for a given year could not be obtained simply by averaging all the clutches reported in that year, because clutch size varies seasonally, and proportionately more late clutches were reported in some years than in others. Clearly, only clutches begun before a certain date could be used for this analysis. This necessarily arbitrary date was taken as twenty days (i.e. four five-day periods) after the peak laying date. Only clutches started prior to this date were included. The advantage of taking a variable, rather than a fixed date, is that it is dependent on, and not biased by, annual variations in the timing of breeding. Also, to eliminate any possible error due to regional differences, only clutches from the south of England (i.e. the south-east and south-west together) were used in this analysis. 58

14 BREEDING BIOLOGY OF CHAFFINCH Table VI shows the mean clutch size in different years for the south of England. Differences between years are very small, the range of variation being o.3 of an egg. An analysis of variance showed that these slight differences are not statistically significant. TABLE VI-THE MEAN CLUTCH SIZE, FOR THE SOUTH OF ENGLAND IN IN- DIVIDUAL YEARS Year Number of clutches Mean clutch size REGIONAL ANALYSIS No geographical differences in mean clutch size were detected between the south-west, south-east, and north of England, when the four regions were analysed separately, but the mean clutch size for Scotland is slightly lower (this difference is not significant). Table VII shows the percentage distribution of various clutch sizes for each region. The figures for the south-west, south-east, and north of England show a surprising constancy in the frequency-distribution of various clutch sizes, but proportionately more clutches of 4 than, 5 eggs were recorded from Scotland. TABLE VII-THE PERCENTAGE DISTRIBUTION OF CLUTCH SIZES FOR THE SOUTH-WEST, SOUTH-EAST, AND NORTH OF ENGLAND AND SCOTLAND Number Clutch size Region of clutches Mean S.E. England % 11% 48% 39% 2% 4.3 S.W. England 710 2% I I % 44% 42% 1% 4.3 N. England 409 1% I2 % 45% 39% 3 % 4.3 Scotland 237 3% 14% 54% 29% 0% 4.1 NESTING SUCCESS There are three principal sources of bias that beset an analysis of nesting success from the fragmentary data given on nest record cards. These are as follows : I) A nest which succeeds for some time is more likely to be found and recorded than one which fails at an early stage. Therefore nests found at an advanced stage in the nesting cycle bias the sample in favour of success, and should be counterbalanced by all those nests which failed at an early age, and so passed unrecorded. As Snow (1955) points out, to obtain a completely unbiased sample, only records of nests found during building should be used, but since these are too few, nests found during the laying and incubation periods have also been included in the present analysis. On this basis data could be extracted from a total of 2,924 cards. 59

15 BIRD STUDY 2) H. Mayer-Gross (pers. comm.) has collected evidence to show that nests most easily noticed by human observers, are also those most likely to be found by predators. It is therefore possible that a greater proportion of nests which pass unnoticed by human observers are successful, than of those more easily found and recorded. This will counterbalance to some extent the bias described under (I) above. Thus, the nest record cards probably represent an intermediate sample; many of the most exposed nests are probably predated before they are found, and many of the best hidden nests are not found at all by human observers. 3) A third source of bias is that it is much easier for an observer to record a failure than a success. Certain success can be recorded only when a nest is visited until fledging, whereas two early visits will often suffice to record a failure. Few nests are visited over the fledging period, and many observers make no further visits after ringing the nestlings, at the eighth or ninth day. In the present analysis, survival of nestlings to nine days has been taken to indicate successful fledging. This greatly increases the number of cards which can be used, but may result in the inclusion of some broods which subsequently fail to fledge. However bias introduced in this way is probably small, because older nestlings generally leave the nest instantly at the approach of a predator, so that some or all of them will escape and survive. By the term `partial failure' is meant that at least one, but not all the eggs in a nest hatch, or nestlings fledge, as the case may be. `Total failure', on the other hand, indicates the complete loss of eggs or nestlings from a nest. OVERALL SUCCESS. I) Including total failures, out of Io,967 eggs laid, 5,979 (59 per cent) hatched, and 4,496 (41 per cent) produced fledged young. 2) Excluding total failures, out of 7,021 eggs laid 6,459 (92 per cent) hatched, and 6,178 (88 per cent) produced fledged young. 3) Of a total of 2,915 clutches started, 56 per cent hatched at least one young, and from 44 per cent, at least one nestling fledged. THE INCIDENCE OF TOTAL FAILURE In Table VIII the percentage of all nests successfully reaching hatching and fledging stages, is listed separately for each year. i) The percentage of all clutches from which at least one egg successfully hatched, varied from 37 per cent in 1951, to 72 per cent in ) The percentage of all nests which hatched young, from which at least one nestling sucessfully fledged, varied from 36 per cent in 196o, to 85 per cent in ) The overall success, that is, the percentage of all clutches started which produce free-flying young, varied from 18 per cent in 196o, to 58 per cent in o

16 BREEDING BIOLOGY OF CHAFFINCH TABLE V1II-ANNUAL VARIATIONS IN THE HATCHING AND FLEDGING SUCCESS Percentage of nests in which at least one egg hatched Percentage of broods in which at least one young fledged Percentage of nests started which produced at least one young Year Number of nests Before 195o o 195o 17o o o o III 72 5o o 161 5o All years In the years 1950, 1951, and 1960, the success rate was significantly below average. In 1951, this low success rate was due mainly to a low hatching success. In the years 1956 and 1957, both the hatching and the fledging success were significantly above average. The success rate in each year depends mainly on the proportion of nests which are subjected to predation, but also to some extent on the weather. One of the chief nest predators is probably the Weasel (Mustela nivalis), and Lack (1958) has suggested that heavy nest predation occurs chiefly in years in which the Weasels main food (mice) are scarce. SEASONAL DIFFERENCES IN THE SUCCESS RATE I) Hatching success. Table IX shows the half-monthly percentages of Chaffinch nests in which some young successfully hatched. There is an TABLE IX-THE PERCENTAGE OF CLUTCHES IN WHICH AT LEAST ONE EGG SUCCESSFULLY HATCHED, AND FROM WHICH AT LEAST ONE YOUNG FLEDGED, FOR EACH HALF-MONTHLY PERIOD Percentage of Percentage of Percentage broods hatched clutches started Period Number of clutches hatched that fledged that fledged AprilI Iol April II May I May II June I o 49 June II 17 (83) (6o) (5o) increase in hatching success as the season progresses, from about 5o per cent at the beginning of April to 7 0 per cent by June. Taking the season as a whole, 56 per cent of the clutches started hatched some young. (These figures do not of course, distinguish between the different causes of failure, of which the chief is nest predation.) In addition, they refer 6 1

17 BIRD STUDY only to nests in which some young hatched, and take no account of the number of eggs which hatch in those clutches which are successful. This aspect is dealt with later. 2) Fledging success. The percentage of nests from which some young fledge (using the criterion of survival as nine days) is shown in Table IX. As in the case of hatching success, fledging success increases throughout the season, when measured as a percentage of all the nests started. However, this is due partly to the increasing hatching success, and when measured as a function of those nests in which some young hatched, a different pattern emerges. Fledging success increases from 75 per cent for nests begun in April, to a peak of over go per cent for nests started during the latter half of May, and then falls off again at the end of the season. Thus in terms of escaping nest predation, the optimal time for young in the nest is towards the end of June, rather later than the peak period. It may be tentatively suggested that broods in the nest in May, when cover is less dense, are more easily lost to predators hunting by sight, and that at the end of the season, when food is probably scarcer, the young are hungrier, and therefore noisier, and more vulnerable to predators hunting by ear. In mid-season however, nests are well hidden by dense vegetation, and food for the young is at a maximum. TABLE X-THE PERCENTAGE OF BROODS OF DIFFERENT SIZES FROM WHICH AT LEAST ONE YOUNG SUCCESSFULLY FLEDGED, ACCORDING TO MONTH IN WHICH THE EGGS WERE LAID. THE DATA FOR APRIL AND MAY HAVE BEEN GROUPED TOGETHER BECAUSE THERE WAS NO SIGNIFICANT DIFFERENCE BETWEEN THEM Brood size 5-6 young 4 young 2-3 young Number Number Number of of of Number success- Percent- Number success- Percent- Number success- Percent- Nests of ful age of ful age of ful age begun in brood., broods success broods broods success broods broods success April and May June 12 7 (58) (68) (74) Table X shows the percentage of broods that produce at least one free-flying young, according to brood size and month. It will be seen that for broods in the nest until about mid-june, the success rate is similar for broods of all sizes. But later in the season, on the limited data available, proportionately more broods of five young are taken by predators, than broods of two or three young. It may be suggested that towards the end of the season, when food is presumably scarcer, the nestlings of large broods are hungrier, and therefore noisier, and more likely to attract the attention of predators, than are young in small broods. The consequence of this differential mortality between broods of different sizes, will be discussed later. 62

18 BREEDING BIOLOGY OF CHAFFINCH 3) The causes of total failure. It has been pointed out above that almost a half of all Chaffinch nests started, fail to produce any free-flying young. The purpose of this section is to review the principal causes of total failure, which are of course, numerically very important. Unfortunately, most observers fail to comment on the cause of destruction, and in most cases, the cause is presumed rather than known with certainty. It is therefore only possible to categorise broadly the causes of failure into predation, desertion, inclement weather, and miscellaneous causes. 1) Predation. Under predation have been included all cases of loss or destruction of eggs, nest, or young. Often desertion follows removal of one or more eggs from a clutch; these cases have been counted as predation. Probably many such cases of desertion are the result of interference by small boys, since it is unlikely that a natural predator would remove one egg from a clutch. 2) Desertion. Failure has been attributed to desertion, only when there was no evidence of predation. Many cases are doubtless due to human interference, or to death of the parent bird, but some cases included under this heading may have been caused by adverse weather. But such cases have as far as possible been included under a separate heading. In addition, some other cases of apparent desertion are probably the result of the bird laying an infertile clutch, and deserting after `outsitting' its normal incubation period. The five definite cases of the whole clutch being infertile are included under miscellaneous causes. 3) Weather. A sudden cold snap sometimes appeared to cause desertion, since it was often reported by several observers simultaneously. Such a cold snap occurred at the end of April, 1953, and caused many birds to desert during incubation. Similarly, rainstorms, or prolonged wet weather frequently caused losses among nestling Chaffinches. 4) Miscellaneous causes. Under this heading are included all other causes of failure, for example, destruction of the nest site, infertility, etc. TABLE XI-APPARENT CAUSES OP TOTAL FAILURE OF EGGS AND YOUNG A. CAUSES OF TOTAL FAILURE OF CLUTCHES Clutches begun in April May June All months Number of nests o Predation.. 63 % 72% (56%) 68% Desertion.. 32% 22% (34%) 27% Weather.. 3% 3% (5%) (3%) Miscellaneous 2% 3% (5%) 2 % B. CAUSES OP TOTAL FAILURE OF BROODS Clutch begun in April May June All months Number of broods 51 6o Predation % 63% 67% -, 6o% Desertion % 30% (25%) 25% Weather.. 22% 5% - (8%) 14% Miscellaneous o% 2% 0% I% E 63

19 BIRD STUDY Taking the season as a whole, predation accounted for about twothirds of the egg losses, and desertion for most of the remainder. About half the cases of egg predation were attributed to small boys, as were many cases of desertion. Predation accounted for about 6o per cent of nestling losses, and desertion for about 25 per cent. In almost all cases, desertion of nestlings was put down to death of a parent bird. Prolonged rain is also an important factor in nestling mortality. For example, a heavy rainstorm which swept the south of England in May 1955, destroyed over half the broods under observation at the time. The percentage loss for nests begun in April attributed to adverse weather, is largely due to this rainstorm, and is therefore probably rather higher than usual. Predation of young increases throughout the season, but the percentage loss due to other causes remains more or less steady. Because natural predation is so rarely observed in the field, very little is known about the nest predators themselves. Natural predators are rarely given with certainty on nest record cards, and those that are, are the ones most easily observed. Of the natural predators listed, the Magpie (Pica pica) accounted for over half the known cases. It is clear that Magpies locate nests mainly by watching the parent birds; most of the cases of predation by Magpies occurred during the nestling stage, when visits to the nest by the parents would be most frequent. Jays (Garrulus glandarius), Grey squirrels (Sciurus carolinensis), and domestic cats (the latter especially in gardens) were frequently recorded as taking nestlings, and there is one case each of predation by Redbacked Shrike (Lanius collurio), Carrion Crow (Corvus corone), and Grass Snake (Natrix natrix). With the exception of the last, all the predators listed are easily seen, but probably Stoats (Mustela erminea) and Weasels accounted for many of the losses. Rats (Rattus sp.) were mentioned only as suspects. There were three cases of parasitism by the Cuckoo (Cuculus canorus), and in at least one of these cases the Cuckoo was successfully reared. In Britain the main laying season of the Chaffinch is over before that of the Cuckoo begins, but in Scandinavia, where the laying seasons of both species coincide, Chaffinches frequently act as hosts to the Cuckoo (Baker, 1942). i) Hatching success in successful nests. It is not always possible to tell from the cards the exact number of eggs in a clutch which hatch, because if a nest is not visited over the hatching period, it is uncertain whether a reduction in the number of young is due to the loss of an egg or of a nestling. Cards not giving precise data on this point have therefore been omitted from analysis. The figures, given in Table XII show that the percentage of eggs that hatch remains about the same throughout the season, but there 64

20 BREEDING BIOLOGY OF CHAFFINCH TABLE XII-THE PERCENTAGE OF EGGS LAID WHICH HATCH, ACCORDING TO CLUTCH SIZE, AND MONTH April May June Number of Percentage Number of Percentage Number of Percentage Clutch size eggs laid hatched eggs laid hatched eggs laid hatched All is a slight decrease in June clutches. This is due mainly to an increase in the proportion of infertile eggs. 2) Fledging success in successful nests. Since they introduce no bias, nests found at a late stage in the nesting cycle may be included here, giving altogether a sample of over 2,50o cards. TABLE XIII-THE MEAN CLUTCH SIZE, MEAN NUMBER OF YOUNG LEAVING THE NEST, AND THE PARTIAL FAILURE IN EACH MONTH. THE PARTIAL FAILURE IS TAKEN AS THE DIFFERENCE BETWEEN THE MEAN CLUTCH SIZE AND THE MEAN NUMBER OF YOUNG FLEDGING. April May June Mean clutch size Mean number fledging (4) Partial failure Table XIII, based on all successful clutches, gives the mean number of young leaving the nest, the mean clutch size, and partial failure in each month. The average number of young successfully leaving the nest decreases steadily throughout the season, but this is largely due to a corresponding decrease in mean clutch size. Thus, partial failure (the difference between clutch size and the number of young fledging), remains more or less steady throughout the season. The commonest causes of partial loss are the occurrence of addled eggs, chipping of eggs, or the displacement of eggs or young from a nest. TABLE XIV-THE MEAN NUMBER OF YOUNG LEAVING THE NEST, ACCORDING TO CLUTCH SIZE, FOR EACH MONTH. 5 eggs 4 eggs 3 eggs Mean Mean Mean Number number Number number Number number of fledged of fledged of fledged clutches per nest clutches per nest clutches per nest April May,8o June 5 (5.0) (2.4) In Table XIV, the mean number of young leaving the nest is given according to. clutch size for each month. There are too few clutches of two and six eggs, and insufficient June clutches, to give meaningful averages., For each clutch size, fewer young fledge from successful nests begun in May than in April. 65

21 BIRD STUDY PRODUCTIVITY Following Snow (1955), the productivity of nests has been taken as the product of the percentage of clutches which reach fledging, and the mean number of young per nest reaching the fledging stage. Table XV shows that productivity of nests increases as the season progresses to about TABLE XV-THE PRODUCTIVITY OF NESTS THROUGHOUT THE BREEDING SEASON. Mean number Percentage of of young clutches reaching leaving each fledging stage successful nest Productivity April May June Mean two young per nest for June nests. This increase occurs despite a concomittant decrease in mean clutch size, low clutch size at the end of the season being more than offset by the greater success of late nests. DISCUSSION In Britain, the Chaffinch lays from two to seven eggs in a clutch, the mean clutch size being 4.3 eggs. The commonest clutch consists of 4 eggs, followed by clutches of 5, and then 3 eggs. For successful nests, on average, clutches of 5 and 6 eggs each give rise to 4.6 young, clutches of 4 eggs to 3.6 young, 3 eggs to 2.7 young, and 2 eggs to 1.9 young. It may also be relevant that from the only brood of 7 young recorded, only 2 young fledged, the rest died before they reached the fledging stage. Thus, in terms of the number of young which survive to the fledging stage, 5 and 6 eggs are the most productive clutch size, and of these clutches of 5 eggs are the most economical, since they involve laying one less egg. Presumably clutch size remains so variable because no one size of clutch is always the most productive. (For a general discussion of the significance of clutch size, see Lack, 1954.) A factor which could influence clutch size in the Chaffinch, is the occurrence of infertile eggs and other partial losses; in almost a half of the clutches recorded on nest record cards, at least one egg failed to hatch. Consider a hypothetical case in which a bird lays a clutch of 5 eggs, of which a happen to be infertile, and assume that three is the optimum brood size. In this way deviations from the normal may persist, at least for short periods. If however, partial loss of a clutch is a regular occurrence, it is advantageous to allow for losses, and lay what would be, if all the eggs hatched, a larger than optimal clutch. The capacity to modify clutch size according to the suitability of conditions for rearing the young, is of course innate, and just as much a product of natural selection, as is the clutch size itself. In the Chaffinch, clutch size decreases from mid-may onwards, despite a concomittant increase in day length. This seasonal decrease in clutch 66

22 BREEDING BIOLOGY OF CHAFFINCH size is therefore primarily adapted to correspond with a food supply which is abundant for a short period in early summer, and then decreases, that is, to the flush of insect food. Towards the end of the season, conditions are likely to be less suitable for rearing large broods of young. However large broods do often occur then, and furthermore the nest records show little or no partial loss of nestlings from starvation. But the fact that smaller clutches persist and are commonest at the end of the season, implies that individuals laying clutches of three eggs at this time are at no selective disadvantage to birds laying larger clutches, and that they leave at least as many ultimate surviving offspring. This may be because young from larger broods suffer proportionately greater mortality after fledging. There is however, another powerful selective factor favouring birds laying smaller clutches at the end of the season. This is the relative occurrence of total predation in broods of different sizes. For much of the season, broods of 3, 4, and 5 young are subjected to predation randomly, but for nests begun in June, broods of 4 and 5 young are taken relatively more frequently than are broods of 2 and 3 young. This is probably because large broods, in the nest during less favourable conditions, are likely to be hungrier, and therefore noisier, and more vulnerable to predation. In other words, by eliminating large broods, natural selection acts so as to favour individuals laying smaller clutches under sub-optimal conditions. SUMMARY I) The egg-laying season of the Chaffinch usually extends from mid- April until mid-june, with a pronounced peak in laying which, on average occurs during the first few days of May. 2) The timing of the breeding season varies between different years, and also between different geographical regions, egg-laying being earliest in the south-west. 3) Air temperatures greatly influence the time of onset, and the pattern of breeding in different years, and by initiating or inhibiting egg-laying, serve to bring the nestling period more nearly to coincide with the peak availability of the insects on which the Chaffinch rears its young. 4) The Chaffinch lays from 2 to 7 eggs in a clutch, the mean clutch size being 4.3 eggs. Mean clutch size remains constant until the end of May and then decreases as the season progresses. It does not vary significantly between different years, or between different regions. 5) The fledging success varies enormously between different years, depending mainly on the proportion of nests which are subjected to predation. The success rate also varies seasonally, the productivity of nests increasing as the season progresses. On average, 56 per cent of the clutches started hatched at least one young, and from 44 per cent, at least one nestling fledged. 67

23 BIRD STUDY ACKNOWLEDGEMENTS Thanks are due primarily to all those members of the B.T.O. who have completed nest record cards for the Chaffinch. Thankful acknowledgement is also due to the current organiser of the nest record scheme, Mr. H. Mayer-Gross for helpful advice in the preparation of this paper, to the late Mr. S. L. B. Lee, for help with statistics, and to Dr. D. Lack for criticising the final draft of the paper. REFERENCES BAKER, E. C. STUART Cuckoo Problems. London. 000DACAE, M. J. and LACx, D Early breeding in Brit. Birds, 52: LACK, D Further notes on clutch and brood size in the Robin. Brit. Birds, 41:98-104, LACK, D The Natural Regulation of Animal Numbers. Oxford. LACK, n A quantitative breeding study of British tits. Ardea, 46: MYRES, M. T. x955. The breeding of the Blackbird, Song Thrush, and Mistle Thrush in Great Britain. Bird Study, 2:2-23. SNOW, D. W The breeding of the Blackbird, Song Thrush, and Mistle Thrush in Great Britain. Bird Study, 2:72-83, SUMMERS-SMITH, D Breeding biology of the Spotted Flycatcher. Brit. Birds, 45: