SHORT COMMUNICATIONS

Do Old Great Tits ForegoBreeding?

ANDR]• A. DHONDT

Departementiologie, niversitairenstelling ntwerpen, -2610Wilrijk,Belgium There is considerable vidence hat young or in-

experienced irds are lesseffectiveparents han old-er individuals,but only limited information s avail-able on the reproductiveperformance f older birds,

especiallyn short-lived pecies.n GreatTits (Parus

major) erritory size increases ith age, reaching

maximumat 4 yr and then decreasingn birds of 5

yr or older (Dhondt 1971). In some casesold Great

Tit males 5 yr or more)do not defenda territoryor raiseyoung n a first brood,but pair with a female

to raise young in her second brood of the season.

Perrins and Moss (1974) found that in Great Tit fe-

males"The birdsof 5 or more yearsof age seem o be becoming essefficientat breeding,rather than curtailing their breeding effort." They showed, n

particular, that older females ay later, have a lower

hatching ate, and produce ledglingswith a lowerchance f survival.The birdsof 5-7 yr did somewhat

poorer han 1-yr-old emales.Webber 1975)showed

that at Oxford the annual survival rates of Great Tits

increased raduallyup to the age of 4 yr and then

decreased teadily or birds of 5 or more years.This

suggestshat all old females will breed, but less ef-

ficiently than when younger, and that older malesoccasionally ill missa reproductive eason.

It is generally extremely difficult to documentnon-

breeding in small passerines. n the populations

studiedat Ghent (Belgium),however, a large pro-portion of GreatTits use nest boxes or roosting n

winter, so hat evidence or the presence f birdscan

be obtained from routine nightly visits to the nest

boxes. present evidence here that Great Tits of both

sexessometimesorego breeding once they reach 5yearsof age. alsoverify for the Ghent populationsthe extent of nonbreedingamong irst-yearbirds, as

has been shown by Bulmer and Perrins (1973).

In the courseof a long-term study of Great Tits atGhent begun n 1959,all nestlingswere ringed and

a variableproportionof breedingadults rappedon

the nest.Adultswere trappedwhen entering he nest

box o feed 1-2-week-oldnestlings.Detailson study areas nd basicnformation n the studycanbe found in Dhondtand Eyckerman1980).Two eveningvisits to all nestboxesweremadeeachwinter,and roosting

birds were trapped.

Of the 962 nestlingsborn between 1959and 1976

that were recaptured fter fledging,417 were caught

at leastonceon the nest.The analysispresented ere

used two subsets of these birds. The first data set con-

sisted of 118 birds (74 males, 44 females) from the

cohorts 1961-1971 that met two criteria: they were

regularly found roostingduring winter, and they

were trapped on the nest n at least half of the pos-sible years.The first conditiongives nformation ona bird's presencen the studyareaeven if it was not

caught on the nest. The second criterion excludes

birds that were difficult to trap or were irregular

breeders in our nest boxes. The 74 males in this set

were trapped 178 times out of a possible 39 (i.e. in74.5%of the possible nstances), nd the 44 females

were trapped 123 times out of a possible144 (85.4%);although this difference was statisticallysignificant

(G-test,Sokaland Rohlf 1969;G = 6.39, 1 df, P < 0.02),

it was an accidental esult of the way the samplewas

drawn together. The relevant question here is, How

did trapping rates differ with age or previous cap-tures? f the age structureof the malesand females

used did not differ, they can be combined nto onesample.

The second data set contained 72 males and 44 fe-

males, not included in the first group, that were

banded as nestlings in the cohorts 1961-1974 and

were known to be alive until at least their second

breedingseason. omeof thesewere not trappedon the nest until 3 yr old; otherswere trapped n theirfirst and secondbreeding seasons ut were neverfound roostingand, therefore,were not included n

the first subsample.My analysiswas basedon the assumption hat the

probabilityof trapping an individual that breeds n

a nest box doesnot vary with age. Van Balen (in litt.)

pointed out that there always s a proportion of un-

trapped parents, because he clutch or brood is de-

sertedbefore trapping s possible.Desertionat an

early stagemay occurmore frequently in very young

or very old individuals. To test this hypothesis,compared he data of two study areas, Maaltepark

and Maria-Middelares, n the period 1966-1969whenthe entire population was color-bandedand knownin sufficientdetail to identify all birds that attempted

to breed.Of 117 breedingpairs hat startedegg-lay-

ing in a first brood, he identity of 115malesand 116femaleswas known. No young fledged from 19 ofthe nests,and the identity and minimum age of all

but ! of these unsuccessful adults were known. Of

90 parentsof age 1 yr or 5 or more years,16 (17.8%)were unsuccessful,omparedwith 21 (14.9%)unsuccessful arentsout of 141aged2-4 yr. This difference

in nest failure between age classeswas not statisti-cally significant G = 0.34, 1 df, P > 0.05).Because ge doesnot seem o affect he probability

of nest ailure, the proportionof individualsactuallycaptured houldreflectdifferencesn the proportion of the birds in eachage class hat breed in the nest

870

October 985] Short ommunications 871

TABLE . Proportion of birds that were trapped on

the nest (first data set).

Age Males Females All

1 51/74 0.69 36/44 0.82 87/118 0.742 58/74 0.78 39/44 0.89 97/118 0.823 33/41 0.80 22/25 0.88 55/66 0.834 24/30 0.80 15/15 1.00 39/45 0.87

5-7 12/20 0.60 11/16 0.69 23/36 0.64

TABLE. Proportionof birds that were trappedon

the nest and were known to be alive until at least

the second reedingseason, ut were not included

in the first data set (second data set).

Age Males Females All

1 32/72 0.44 28/44 0.64 60/116 0.532 51/72 0.71 33/44 0.75 84/116 0.72

boxes. n the Ghent studyareas,where nest-box en-sity is high, practicallyall Great Tits use nest boxesfor breeding,and very few, if any, breed in natural

cavities.

A higher proportionof malesand femalesaged 2-

4 yr was rapped on the nest han of younger or olderbirds (Table 1). Among malesnone of the differences

between age-classes as significant,but among fe-

males the birds of 5 or more years were captured n

significantly ower proportions han thoseof 2-4 yr

(G = 4.57, 1 df, P < 0.05). Because he age structureof the two sexes n the sample did not differ (G =

1.12, 4 df, P > 0.50), ! combined hem for analysis.Two- to 4-yr-oldGreatTitswere trappedsignificantly

more often on the nest than were birds of 1 yr (G =4.44, 1 df, P < 0.05) or of 5 or more years (G = 6.66,

1 df, P < 0.01).

The seconddata set was used o evaluate he prob-

ability that birds aged 1 and 2 yr would be capturedon the nest (Table 2). Two-year-old males were

trapped significantly more often on the nest thanwhen they were 1 yr old (G = 10.41, 1 df, P < 0.01),

but this was not the case for females (G = 1.34, 1 df,P > 0.20). When I combined both data sets to in-

creasesample size, these results were confirmed:

maleswere trappedmore often when 2 yr old (G =

10.37, 1 df, P < 0.01), but females were not (G = 2.08,1 df, P > 0.10).

How far do differences n the proportion of cap-tures with age reflect differences n the proportion

of breeders?My data show that nest failure, beforeparentscould be trapped feeding nestlings,was notmore frequent among young or very old parents.

However, birds that were trapped on the nest once

when feeding young were more difficult to trap a

second ime in the sameseason pers. obs.). ndivid-

uals hat were very difficult to trap altogetherwere

excludedby the way the first data set was constitut-ed. If experienced irds alsowould becomemore dif-

ficult to trap on the nest because hey had been

trapped in a previous season, nd assuming hat all

birds breed from the age of 2 onwards,one would

expect the proportion of birds trapped to decreasegradually with age. This was not the case Table 1).

I thereforeassumed hat the proportion of capturesreflected he proportionof breedersn all ageclasses,

and, therefore, old birds of both sexes did not breed

each year. Similarly, some yearling males did notbreed,althoughall yearling femalesdid.

What proportions f young and old birds breed?Assuminghatall birdsaged2-4 yr breed,and know-

ing that in theseage classes 9.3%of the malesand

90.5%of the females (seeTable 1) were captured, one

can calculate he proportionof breeders n the other

ageclasses.hesecalculationsndicate hat 75.6%of

the malesand 76.0%of the females ged5-7 yr would actuallybreed.When the calculationswere repeated

for yearlingmales, he percentagesound were 87%

for the first data set, 63% for the second data set, and

76% for the combined data. This is close to the esti-

mate of 66%made by Bulmer and Perrins (1973) for

the WythamGreatTits at Oxford.The difference e-

tween the two data sets n the proportion of year-

lingsbreedingwasstatisticallyignificant:n the first

set 64/74 (86.5%) would have bred, compared with

45/72 (62.5%) in the second set (G = 11.41, 1 df, P <

0.01). This suggestshat birds in the two subsets re

not equivalent n termsof breedingperformance. hemain differencebetween he two groupswas that all

birds n the first groupusednestboxes or roosting

from their first winter onwards, whereas birds in the

secondgroup did not. This suggestshat males hatuse nest boxes or roosting n their first winter aremore ikely to obtaina breeding erritory and a mate.GreatTits roostduring winter in a domicile Kluyver1951;pers. obs.).This domicilenormally overlaps o a large extent with both the autumn and the spring

territory (Dhondt 1971). Furthermore, more domi-

nant ndividualshavea greater hance f being found

roosting n nest boxesduring winter than lessdom-inant individuals (Kluyver 1957, Drent 1983). These

observations upport the conclusion hat the more

dominant yearling Great Tit males have a higher

chanceof breeding n their firstbreedingseason. he

samemay be true for old malesbecause have ob-

served hat someof the old malesbred only a second

brood, without having defended a spring territory.These results emphasizeprevious observationsof

the effectsof old age in Great Tits on weight andwinglength (van Balen 1967), erritory size (Dhondt1971),breedingperformancePerrinsand Moss1974),

and annual survival rate (Webber 1975). It should be

emphasized,however, that only very few individu-