Will Great Tits become Little Tits?

BiologicdJoumal @the Linnean Society, 1 I: 289-294. With 1 figure May 1979 Will Great Tits become Little Tits? ANDRE A. DHONDT, ROMAN EYCKERMAN AND JAN HUBLE Departement Biologic, Universitaire Instelling Antwerpen, B-2610 Wilrijk, Belgium Laboratorium voor Oecologie der Dieren, Zoogeograjie en Natuurbehoud, Rijkjksuniverjksiteit Gent, B-9000 Ghent, Belgium Accepted for publication December 1978 The mean size of Great Tits Parus mqor in our study at Ghent shows a significant decreasing trend in the period 1962 to 1975. Since it has been shown that in the Great Tit the heritability of tarsus length, a measure for size, is rather high, we believe the observed trend to be a directional microevolutionary change. After investigating several possible causes for the size decrease, we suggest that the observed change is caused by a shift in the equilibrium between selective pressures favouring large and small individuals. Before our study, breeding numbers of Great Tits were limited through the lack of suitable nest-sites. We think that more of the large males could reproduce in that situation. When we provided nestboxes in surplus this selective pressure was relaxed, and since small females must produce more surviving offspring, this further shifts the balance. We argue that during the evolutionary history of the Great Tit there was no lack of suitable nestsites in natural habitats, that by providing nest boxes we restore the natural situation, and that therefore the decrease in size should level off. CONTENTS Introduction....................... 289 Material and methods..................... 290 Results......................... 291 Discussion........................ 29 I Conclusion........................ 293 Acknowledgements..................... 293 References........................ 294 INTRODUCTION Directional changes in a quantitative character of individuals in a population are poorly documented in field situations, although numerous examples of character displacement show they must occur. We report here that in our long term study of the Great Tit Parus major the mean size of individuals has decreased over the years, and will speculate as to the possible causes for this change. We will argue that a directional micro-evolutionary change has occurred. 289 0024-4066/79/090289-06/$02.00/0 0 1979 The Linnean Society of London

290 A. A. DHONDT ET AL. MATERIAL AND METHODS Since 1959 Great Tits have been studied at Ghent in a number of wooded study areas (for details of the study areas, see Dhondt & HublC, 1968). At the start of the study a surplus of nestboxes was provided as breeding and roosting sites. Routinely all nestlings were ringed. Many of these were later recovered and since 1962 their wing length was measured by laying the closed left wing on a flat ruler, and sliding the wing, without deforming it, against a stop at the zero mark. HublC's measurements are consistently larger than those of Eyckerman and Dhondt, but no difference was found between the latter two observers. In the analysis of the wing length data we will use Hublk's (JH) data separately, and combine those of Dhondt and Eyckerman (ADRE). Wing length varies throughout the life of a bird: between moults wing length decreases through abrasion, especially during the breeding season. Wing length increases with age, after moult (Van Balen, 1967; extensive own observations). We have therefore limited our sample to juveniles recaptured during their first winter (October-March). We thus measured the feathers that grew while the young were in the nest, since in the partial postjuvenile moult of the Great Tit the primaries are not replaced (Flegg & Cox, 1969; Dhondt, 1975). In four years adults were weighed when trapped on nestlings. Since adult weight varies both daily and seasonally (Van Balen, 1967; extensive own 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 '62 '64 '66 '68 '70 '72 '74 Year Figure I. Wing length and year class of Great Tits Pam major at Ghent. The mean wing length (f 1 standard deviation) of first winter birds is plotted against the year in which they were born. Males (open symbols) and females (filled symbols) are protected separately. Friends represent means based on fewer than 5 individuals. The upper half of the figure is based on JH measurements, the lower half.on measurements by AD and RE. All four least square regression lines have significant negative slopes (see text) indicating that the wing length of the birds in the population has decreased during the study period.

WILL GREAT TITS BECOME LITTLE TITS? 29 1 observations) we limited our sample to adults trapped on their first brood of the season, and only if trapped on young 10 to 15 days old. We have used data from two of our study areas only, e.g., Maria-Middelares, a mixed deciduous park, and Hutsepot, a beech wood. All weights were corrected for time of day, by a correction factor of 0.08 g/h, calculated from all our data. RESULTS Mean wing length per year class, and separately for males and females and per observer is shown in Fig. 1. The reduction in size over the years is highly significant for all four calculated least squares regression lines (all data, not the means, were used in these calculations). The slopes of these four regression lines are all negative: measured by JH males (1962-1970): b=-0.19, ts,= 3.61 P < 0.001 females (1963-1970): b=-0.19, t,,=2.43 P(O.02 measured by AD, RE males ( 1966-1 975) : b = -0.16, t,,, = 3.61 P < 0.01 females (1966-1975): b=-0.17, t,,=2.59 P< 0.02 Mean weight of adults on nestlings, although available only for four years, shows a similar decrease over the years (Table 1). It too is statistically significant, as a one way analysis of variance yields following values : Males Fa, = 6.254, P < 0.001; females Fs, 82 = 5.911, P < 0.001. Both regression coefficients are negative and different from zero : males b = -0.12 (P < 0.0 1 ; females b = -0.07 (P< 0.01). Table 1. Adult weight and year of breeding. In both sexes adults weight has significantly decreased over the years (See text) Males Females Year Number Mean S.E. Number Mean S.E. 1968 15 19.27 0.241 16 18.66 0.192 1969 20 18.76 0.219 27 17.88 0.185 1976 18 18.58 0.254 20 17.68 0.165 1977 31 17.89 0.214 23 17.64 0.150 DISCUSSION The mean size of individuals in a population should be determined by the equilibrium that is reached when selective forces tending to increase or decrease the size are in balance. If selective forces change in a permanent way, the mean size should evolve towards a new equilibrium. Two questions arise here: (i) is the heritability of size in the Great Tit established, so that a change in selective forces can have an effect? and (ii) can we show that a change of selective pressures is likely to have occurred in our population? Garnett (1976) has shown that in the Oxford population of the Great Tit about three quarters of variation in tarsus length, an indicator of body size, is caused by additive genetic variance. All body measurements, such as wing length, tail

292 A. A. DHONDT ET AL length, tarsus length and weight are highly intercorrelated (Eyckerman, personal observation). The answer to the first question is thus positive and the first necessary condition for the observed change to be a directional microevolutionary change is fullfilled. The answer to the second question mus: be largely speculative, since we have performed no experiments to test them. We will however formulate a testable hypothesis, so that our ideas can be tested in the field. A non exhaustive list of possible changes over the years is as follows : ( 1) the carrying capacity of the habitat has changed; (2) predator pressure has changed; (3) interspecific competition has changed; (4) intraspecific competition has changed. Hypothesis 1, a change in the carrying capacity of the environment can be split into two subhypotheses : (a) The amount of food has changed. We know that Great Tits react to an increase in the carrying capacity of their environment by an increase in breeding density (Kluyver, 1951, 1971). A necessary condition for this hypothesis is that we find a significant trend in breeding density. This is not present, and we can therefore reject this hypothesis. (b) Temperature has changed over the years. The temperature data from Ghent do not show a signficant trend over the years in the breeding season, nor in winter. We can therefore reject this hypothesis. Hypothesis 2. An anonymous referee suggested that predator pressure (Sparrowhawks) has increased, and that predators would selectively feed on larger individuals, causing thereby the decrease in size. In our areas predators are very rare and there is no indication that they are becoming more common. Even if predators were increasing in numbers and if they selectively ate larger individuals, the observed change in size would still be an example of a directional micro-evolutionary change. Hypothesis 3. We know that interspecific competition between the Great and the Blue Tit is an important factor in our study areas. This is true both during the breeding season (Dhondt, 1977) and in winter (Dhondt 8c Eyckerman, in prep.). One would expect that, as a result of this competition the two species would become more unlike, and that therefore the Great Tit, being the larger of the two species, would become even larger. It is hard to understand that the result of this interspecific interaction would be that Great Tits became smaller, and we therefore reject this hypothesis. Hypothesis 4. Putting up a surplus of nestboxes in a managed wood results in a doubling or tripling of the breeding density in the Great Tit (Dhondt, Eyckerman & Hublt!, in preparation). Intraspceific competition for nestholes is suddenly relaxed, but intraspecific competition for territories and food is suddenly increased. There is therefore a marked change in intraspecific competition and this hypothesis must be explored in more detail. In areas without nestboxes only some of the birds can breed. The more dominant males are likely to succeed in obtaining the nest-sites available and the less dominant males should be excluded from breeding. In many species dominance is linked to size (Wilson, 1975) and in the Great Tit larger birds are likewise dominant over smaller ones (Garnett, 1976; own observations). Thus we expect that in areas devoid of nestboxes, there will be a strong selective pressure

WILL GREAT TITS BECOME LIlTLE TITS? 295 favouring large males, for these will have a higher chance to reproduce. When nestboxes are provided in surplus, most, probably all males can breed in our areas (own observations). Thus the selective pressure favouring large males will be relaxed, as nest sites are no longer a limiting factor. Since small males will then reproduce, the mean size of the individuals in the population will decrease. There might be other factors favouring small individuals. Thus smaller females, laying smaller eggs, need less energy to form a clutch, and lay earlier in the season (Jones, 1972). Young that fledge earlier in the season survive better (Perrins, 1965; Dhondt, 1971). Therefore smaller females must have a higher fitness compared to larger ones. Providing nestboxes, in this view, should result in a shift of the equilibrium between selective pressures favouring large males (that had a better chance to reproduce when nest-sites were limiting) and small females (that produce more surviving young). This change could lead to the directional micro-evolutionary trend observed. 0 ther factors (winter temperature, food distribution and abundance,...i that vary between years, could lead to changes in selective pressures that are different from year to year and could therefore not cause the change in size observed. These would rather be responsible for variations around the trend. This last hypothesis is testable in the field. Two Great Tit populations could be compared. In one of these nestboxes should be added, and our hypothesis predicts that in this population the mean size of individuals should decrease, but not in the other one. CONCLUSION It is im ortant to ask whether the increased population density, sue to the addition o P nestboxes, and possibly the driving force behind the oherved change in size, creates an artificial situation or would rather be a return to a more natural situation. Perrins ( 1965) thinks that in primaeval woodlands surplus nest-sites must have been available. We agree for the following reason. If Great Tits had, during their evolutionary history, always been limited by nest-sites, they would have developed an active nest-site defence. If nest-sites were abundant, but another resource limiting, they would have developed a territorial system to defend the limiting resource, regardless of the number of nest-sites. We know that, in optimal habitats saturated with nest boxes, Great Tit breeding density is limited, at least in some years, by spring territorial behaviour (Dhondt, 1971; Krebs, 1977). Great Tit song is varied and has a repellent effect on potential settlers to the area (Krebs, 1977). We believe that this behaviour could not have developed unless some resource, such as space (or the food it contains), rather than nest-sites, had been the limiting factor during the evolutionary history of the species. If this reasoning is correct, the observed decline in Great Tit size around Ghent, must be a return to a more natural situation, that existed before nest-sites were removed with dead trees. The decrease in size will thus level off and Great Tits will not end up by becoming Little Tits. ACKNOWLEDGEMENTS We are grateful to Miss J. Splingaer, who wrote the computer programs. Many people found time to discuss our ideas or read earlier drafts of the manuscript.

294 A. A. DHONDTETAL We therefore thank Malte Anderson, Judy Meyers, Charlie Krebs, Jamie Smith, Bob Montgomerie, T. R. E. Southwood, Bill Reed and Bob Gauldie. Part of this work was supported by a grant of the Belgian National Foundation for Scientific Research (N.F.W.O.) to A. Dhondt. REFERENCES DHONDT, A. A., 1971. The regulation of numbers in Belgian populations of Great Tits. Proceedings of the Advanced Study Institute on Dynamics of Numbers in Populations (Ocsterbeek, 1970): 523-547. Wageningen: Pudoc. DHONDT, A. A., 1973. Post juvenile and postnuptial moult in a Belgian population of Great Tits, Parus map, withsomedataoncaptivebirds. DeCiervalk, 63: 187-210. DHONDT. A. A., 1977. interspecific competition between Great and Blue Tit. Nature, 268: 521-523. DHONDT, A. A. & HUBLE, J., 1968. Fledging date and sex in relation to dispersal of young Great Tits. Bird Study, 15: 127-134. FLEGG, J. J. M. & COX, C. J., 1969. The moult of British Blue Tit and Great Tit populations. Bird Study, 16: 147-157. GARNETT, M., 1976. Some aspects ofbody-size in the Great Tit. D. Phil. Thesis, Oxford University. JONES, P. J,, 1972. Food as a proximate factor regulating the breeding season of the Great Tit (Parus major). Proceedings ofthe X Vth International Ornithological Congress: 657-658. Leiden: Brill. KLUYVER, H. N., 1951. The population ecology of the Great Tit, Pam m. major L Ardea, 39: 1-135. KLUYVER, H. N., 197 1. Regulation of numbers in populations of Great Tits (Parus m. mqbr). Proceedings ofthe Advanced Study Institute on Dynamics of Numbers in Populations (Oosterbeek, 1970): 507-523. Wageningen: Pudoc. KREBS, J. R., 1977. Song and territory in the Great Tit Parus mqor. Euolutionaly Ecology: 47-62. London: Macmillan. PERRINS, C. M., 1965. Population fluctuation and clutch size in the Great Tit, Parus major L.Journal ofanimal Ecology, 34: 601-647. VAN BALEN, J. H., 1967. The significance of variations in body weight and wing length in the Great Tit, Parus major. Ardea, 55: 1-59. WILSON, E. O., 1975. Sociobiology. Cambridge, Mass.: Belknap.