• Ashmole's hypothesis;
  • clutch size–lay date relationships;
  • introduced birds;
  • latitudinal gradients;
  • seasonality


  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References
  • 1
    Clutch sizes generally increase with latitude, and are smaller at southern latitudes compared with equivalent northern ones.
  • 2
    Descriptions of such patterns and attempts to identify their causal mechanisms are complicated as different species, with different ecological traits are often compared in different regions. We reduce such problems by using the introduction of 11 passerine species from the UK to New Zealand as a natural experiment to explore interspecific geographical variation in clutch size.
  • 3
    Nine species have significantly smaller clutches in New Zealand than the UK. Seasonality, measured both by climate and how birds respond to variation in resource availability, is also lower in New Zealand. Comparing across species, the magnitude of clutch size change is unrelated to the magnitude of reduced seasonality that each species experiences.
  • 4
    Such observations are partly compatible with Ashmole's hypothesis that areas with high seasonality have large clutch sizes (higher winter mortality results in a breeding population that is significantly lower than the environment's carrying capacity, and hence in extra resources for rearing chicks). However, additional data on seasonal changes in resource availability and population densities, combined with comparative data on survival and nest predation rates, are required to evaluate fully the mechanisms generating smaller clutches in the southern hemisphere. We discuss the potential determinants of geographical variation in the patterns of temporal variation in clutch size.


  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

Spatial variation in avian clutch size exhibits two striking patterns. First, clutch size increases with latitude both within and among species (Lack 1947, 1948; Skutch 1949; Dunn et al. 2000; Cardillo 2002). Second, the slope of the clutch size–latitude relationship is shallower in the southern hemisphere, so that clutches laid in the south are generally smaller than clutches laid at equivalent latitudes in the north (Moreau 1944; Yom-Tov 1987; Yom-Tov, Christie & Iglesias 1994; Martin 1996; Rowley & Russell 1991; van Zyl 1999). These relationships have been well documented and several mechanisms have been advanced to explain this geographic variation, but the nature of the underlying causal mechanisms remains elusive. Indeed, accounting for such variation has been described ‘as exciting and challenging a problem as exists in biology’ (Ricklefs 2000).

Ashmole's hypothesis states that clutch size increases with latitude as a consequence of an increase in seasonality. At higher latitudes, that is more seasonal environments, colder winters result in reduced carrying capacity and increased winter mortality, lowering the number of individuals that survive to enter the breeding season. Because there are fewer survivors to take advantage of the flush of resources at the start of the breeding season, there is higher per capita resource availability allowing larger clutches to be raised (Ashmole 1963). Greater seasonality in the northern relative to the southern hemisphere (e.g. Gaston & Chown 1999) could likewise explain why birds in the northern hemisphere tend to lay larger clutches than those at equivalent southern latitudes. Ashmole's hypothesis has received support from studies describing the relationship between clutch size and seasonality within tropical habitats (Lack & Moreau 1965) and linking clutch size variation to the ratio of summer to winter resource abundance measured indirectly as the summer/winter ratio of actual evapotranspiration (Ricklefs 1980; Koenig 1984; Skutch 1985; Young 1994; Dunn et al. 2000).

There are two significant problems in trying to analyse causes of geographical variation in clutch size. First, because the pattern is expressed over a wide latitudinal range, studies often compare different sets of species in different latitudinal regions. This complicates interpretation because different species can vary in ways that may confound the apparent relationship between clutch size and latitude, such as in habitat use and susceptibility to predation. This problem is reduced by comparing closely related species from the same genus or family (Martin et al. 2000a; Cardillo 2002), but such species may still vary in ways that confound comparative analyses. Second, most studies have compared clutch sizes in species from tropical and northern temperate regions (but see Martin et al. 2000a). Consequently, even when the same species occurs in both regions, interpretation is complicated because factors such as habitat and day-length covary with geographic location, making these effects difficult to separate from other factors of interest, such as seasonality.

One way to circumvent these problems is to compare populations of the same species that occupy similar habitats at similar latitudes in the northern and southern hemispheres, thereby removing the confounding effects of phylogeny, latitude and habitat. Few species have natural distributions of this type, but the distributions of many bird species have been extended by human introductions (Duncan, Blackburn & Sol 2003). Here, we explore the causes of geographical variation in clutch size by comparing populations of 11 passerine species that occupy similar habitats at equivalent temperate latitudes in two locations in the northern and southern hemispheres: United Kingdom (UK), where the species are native, and New Zealand, where they have been introduced. In New Zealand the 11 species mainly occupy farmland and urban habitats, which are very similar to the habitats they occupy in the UK because agricultural practices are comparable and because the majority of plant and animal species that dominate farmland and urban habitats in New Zealand were introduced from the UK. We test whether clutch sizes are smaller in the southern (New Zealand) than the northern hemisphere (UK) and if such differences are related to differences in seasonality, as predicted by Ashmole's hypothesis.

Studies that have tested Ashmole's hypothesis using summer/winter ratios of actual evapotranspiration as a proxy for seasonality assume that seasonal variation in the availability of resources limiting bird reproduction reflect variation in the total resource base. The validity of this assumption is uncertain as species typically use a limited subset of resources whose availability may not mirror that of total plant productivity. Moreover, availability may be influenced by many factors other than plant productivity, such as day-length (Sanz et al. 2000; Hill et al. 2003) and habitat structure (Butler & Gillings 2004; Whittingham & Evans 2004; Butler et al. 2005), which are likely to have a large influence on the food resources birds can acquire.

In this paper we adopt an alternative approach to measuring seasonal variation in resource availability. Specifically, we use two measures of how birds respond to seasonality that reflect the magnitude of annual variation in resource availability: (i) temporal variation in clutch size and (ii) breeding season length (Crick, Gibbons & Magrath 1993; Fig. 1). Empirical data strongly suggest that, for our focal species and others, temporal variation in clutch size is more marked and breeding season length is shorter in higher latitude environments with greater annual seasonal variation (Cramp, Simmons & Perrins 1977–94; Dhondt, Kast & Allen 2002; Cooper, Hochachka & Dhondt 2005; Gil-Delgado et al. 2005). Given that species vary in their response to a given level of seasonality, our approach allows these differences to be related to interspecific variation in geographical patterns in clutch size. If Ashmole's hypothesis is correct, we predict that smaller average clutch sizes in the southern hemisphere will be associated with less seasonal conditions, reflected in longer breeding seasons and less temporal variation in clutch size throughout that season.


Figure 1. A model of temporal change in clutch size for species capable of laying multiple clutches per breeding season (see Crick et al. 1993) in relatively seasonal (solid curve) and aseasonal (dashed curve) environments. The curves show changes in per capita resource availability throughout a single year, and concomitant changes in average clutch size. Breeding commences when resource availability exceeds the threshold indicated by the dotted line. In more seasonal environments (solid curve), greater winter mortality, caused by lower per capita resource availability, results in smaller breeding populations at the start of the breeding season and this, coupled with a more marked increase in total resource availability results in higher per capita resource availability, allowing larger clutches to be provisioned. Clutch size (CS) then varies in proportion to per capita resource availability throughout the breeding season, and so follows the portion of the curves above the dotted line. The arrowed lines indicate the ratio of maximum to minimum clutch size. CSR thus measures how species respond to variation in per capita resource abundance during the breeding season. The magnitude of annual variation in resource availability is positively correlated with variation in resource availability within the breeding season. The amount of each curve above the dashed line indicates breeding season length, which is expected to be longer in less seasonal environments.

Download figure to PowerPoint

Materials and methods

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References


We studied 11 passerine species native to the UK that are established in New Zealand (listed in Table 1). Founder populations of each species were mainly introduced from the UK between 1862 and 1885 (Thomson 1922). For each of these species we obtained breeding data from national nest record card schemes organized by the British Trust for Ornithology (BTO), described by Crick, Baillie & Leech (2003), and the Ornithological Society of New Zealand (OSNZ), described by Robertson (1986). Our UK data span the years 1923–2000 (39 595 nest records) and the New Zealand data cover the period 1925–2002 (3350 nest records). In both cases the vast majority of data are from more recent decades.

Table 1.  Comparison of clutch sizes in 11 bird species in their native United Kingdom (UK) and introduced New Zealand (NZ) ranges. Mean clutch size values are presented ± SE, with sample size in parentheses. Two-sample t-values are given with the degrees of freedom as the subscript
SpeciesClutch size NZClutch size UKClutch size NZ−UKt-valueP-value
  • µ

    t-tests conducted using the Satterthwaite method for unequal variances.

  • e

    t-tests conducted using the pooled method for equal variances.

Fringilla coelebs3·68 ± 0·08 (140)4·23 ± 0·01 (4942)−0·55t5080 = 7·58e< 0·0001
Turdus merula3·32 ± 0·03 (752)3·83 ± 0·01 (3615)−0·51t1141 = 17·15µ< 0·0001
Sturnus vulgaris4·15 ± 0·06 (318)4·51 ± 0·02 (3958)−0·36t380 = 6·26µ< 0·0001
Passer domesticus3·74 ± 0·05 (370)4·09 ± 0·02 (3394)−0·35t454 = 7·21µ< 0·0001
Alauda arvensis3·14 ± 0·07 (107)3·46 ± 0·02 (1942)−0·32t2047 = 4·58e< 0·0001
Carduelis chloris4·40 ± 0·06 (148)4·72 ± 0·01 (5094)−0·32t5240 = 4·84e< 0·0001
T. philomelos3·74 ± 0·02 (782)4·05 ± 0·01 (7388)−0·31t1017 = 12·47µ< 0·0001
C. carduelis4·55 ± 0·04 (362)4·72 ± 0·03 (916)−0·17t781 = 3·53µ< 0·0001
Prunella modularis3·83 ± 0·05 (228)3·97 ± 0·01 (5906)−0·14t6132 = 2·59e  0·0097
C. flammea4·29 ± 0·09 (78)4·41 ± 0·07 (223)−0·12t163 = 1·00µ  0·32
Emberiza citrinella3·34 ± 0·09 (65)3·44 ± 0·02 (2217)−0·10t2280 = 1·09e  0·27

We calculated clutch size and clutch initiation date (termed first egg date) using standardized procedures developed by the BTO, using only data from nests visited more than once (Crick et al. 2003). Calculations required knowledge of the maximum possible delay during egg-laying, i.e. the refractory period between the laying of successive eggs, maximum and minimum incubation periods and maximum hatching intervals. We set these variables equal to the most extreme values recorded in either the BTO data or in information on the species’ breeding biology in New Zealand (Heather & Robertson 1997).

Nest contents were rarely recorded on a daily basis, so estimates of maximum and minimum first egg dates were obtained for most clutches. First egg dates were calculated as the mean of the maximum and minimum first egg dates, taking leap years into account and using the winter solstice as day zero. We discarded records where the maximum and minimum first egg dates differed by more than 10 days; first egg dates were thus accurate to within 5 days.

statistical analysis

To determine if clutch sizes of each species differed between countries we first used two-tailed t-tests. We then used multiple regression to construct minimal adequate models (MAMs) of clutch size in relation to country, while taking first egg date (and its squared term), altitude, year, longitude and absolute latitude into account. First order interactions between country and the other main effects were also included as predictors. The inclusion of absolute latitude as a predictor takes the latitudinal difference between the UK and New Zealand into account, and thus a significant effect of country cannot arise from this latitudinal difference. Longitude and its interaction with country were included as predictors to account for broad-scale geographical variation in factors that may influence clutch size, for example the west coast of New Zealand is much wetter than the east. When interactions between longitude and country were insignificant, both the interaction term and longitude were removed from the model because strong collinearity between longitude and country would reduce our ability to detect an effect of country if longitude was retained as a main effect. Altitude and first egg date were included as predictors as clutch size often responds strongly to these variables (Crick et al. 1993).

Within each country, nests were not evenly distributed in space or time but were clustered within years and regions. Clutch sizes may thus show spatial and temporal autocorrelation, invalidating the assumption of independence and generating a greater frequency of type I errors. To deal with this, we fitted linear mixed models that included year and region as random effects. Within each country, these models assumed a common correlation between nests in the same region or year, and zero correlation, or independence, between nests from different regions or years. Our spatial clustering variables comprised New Zealand's 268 bioregions, which are classified by soil type, climate and habitat (McEwan 1987), and the 103 UK counties/regions that, although being administrative units, divide the country into areas of comparable habitat and climate. A conservative approach was adopted to control for the effects of the clustering variables; they were retained even if their effects were not statistically significant (i.e. P > 0·05). For each species, the MAM that best explained variation in clutch size was obtained by backwards deletion from a full model. Interactions between country and first egg date (and its squared term) are of particular interest because significant interactions imply that birds in different countries have different patterns of clutch size variation throughout the breeding season.

To further investigate country differences in each species’ breeding biology we obtained two measures of how birds respond to seasonality: (i) a measure of temporal variation in clutch size and (ii) breeding season length. The clutch size of multibrooded species, such as the British populations of all 11 species in this study, exhibits a mid-season peak that reflects variation in resource availability (Crick et al. 1993; Fig. 1). In order to raise more than one clutch, breeding is initiated before conditions are optimum. Clutch sizes are thus initially small, but increase as conditions improve and then decline when conditions deteriorate at the end of the breeding season. In less seasonal environments per capita resource availability should increase less markedly from the start to the peak of the breeding season, and decline less markedly at the end of the breeding season, because the changes in absolute resource availability and population abundance are more muted between seasons. Similarly if species are single brooded and exhibit a linear decline in clutch size from the start to the end of the breeding season, the magnitude of this decline should be less marked in less seasonal environments. At the extreme there will be no change in relative resource availability in a completely aseasonal environment. Therefore, although other factors may influence temporal variation in clutch size (Christians, Evanson & Aiken 2001) birds should exhibit less temporal variation in clutch size throughout the breeding season in less seasonal environments. In addition, breeding season length is typically longer in less seasonal environments because suitable breeding conditions occur earlier in the spring and are curtailed later. Empirical data support these assumptions (Cramp et al. 1977–94; Young 1994; Dhondt et al. 2002; Cooper et al. 2005; Gil-Delgado et al. 2005).

For each species, we used the parameter estimates from the MAMs to obtain equations relating clutch size to first egg date for each country. We then used these equations to quantify temporal variation in clutch size as the ratio of the maximum and minimum clutch sizes observed during the season (CSR). When other variables, such as altitude, were retained in the MAM, we calculated CSR using the average value for that variable. The proportional change in CSR between the UK and New Zealand was then expressed as (CSRNZ − CSRUK)/CSRUK.

We used two indices of breeding season length. First, we calculated the number of days between the 1% and 99% quantiles of first egg dates (interval index). This reduces the influence of outliers and thus interannual variation in the timing of the breeding season length. To correct for the larger number of UK cards we randomly selected y nests, where y is the number of nests that fall between the 1% and 99% quantiles in New Zealand, and calculated the season length from this sample. We repeated this procedure 5000 times and used the average as an estimate of the breeding season length in the UK. Second, following MacArthur (1964), we calculated the number of equally good 14-day periods for nesting using the expression exp[–ΣPi(logePi)] (MacArthur index). Pi is the proportion of nests found during the ith 14-day period. The proportional change in season length between the UK and New Zealand was expressed as (season lengthNZ − season lengthUK)/season lengthUK.

We tested for differences between the two countries using one-sample t-tests to determine if proportional changes in season length and CSR differed significantly from zero. Finally, to assess the extent to which changes in seasonality may explain changes in clutch size among species, we regressed proportional change in clutch size against proportional change in seasonality (CSR and the two measures of breeding season length) between countries. To ensure that these last relationships were not confounded by phylogeny, we analysed the relationships both across species and within taxa using the CAIC program, version 2·6·9 (Purvis & Rambaut 1995). For the phylogenetic analyses, we classified species according to the phylogeny of Sibley & Ahlquist (1990), assumed that all branches in the phylogeny were of equal length, and assessed the robustness of the resulting relationships to phylogenetic and statistical assumptions of the method using standard CAIC diagnostics.


  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

Nine of the 11 species laid, on average, significantly smaller clutches in New Zealand than in the UK (Table 1). The two exceptions, C. flammea and E. citrinella, also followed this trend. Averaged across all species, clutch sizes in New Zealand were reduced by 8% (0·3 of an egg). Interactions between first egg date and country were significant in the MAMs for eight species, demonstrating that, for these species, the pattern of temporal variation in clutch size differed between countries (Table 2, Fig. 2). These interaction terms were not significant for P. domesticus, C. flammea and E. citrinella.

Table 2.  Fixed-effect predictors of clutch size retained in mixed effect minimal adequate models (MAMs). Values represent F ratios with one numerator degrees of freedom, ddf = denominator degrees of freedom, feg = first egg date
SpeciesddfCountryfegfeg2feg × Countryfeg2 × CountryLatitudeLatitude × CountryAltitudeAltitude × Country
  • *

    P < 0·05,

  • **

    P < 0·01,

  • ***

    P < 0·001,

  • ****

    P < 0·0001.

Fringilla coelebs3212 12·9***  1·5 92·9**** 14·9**** 15·7****    
Turdus merula3779 43·7****139·9****116·6**** 56·7**** 54·3****  0·611·6***
Sturnus vulgaris3206 15·9**** 24·4**** 37·9**** 15·4****     
Passer domesticus1925  4·7* 13·1*** 15·2****      
Alauda arvensis1315  1·4  8·9**  7·0** 13·8*** 14·2***  2·7 4·9*
Carduelis chloris2512  4·3* 19·5****214·2****  4·6*  4·3*    
T. philomelos5920185·4****221·6****363·6****224·8****209·3****  7·3** 
C. carduelis1000  1·0 47·0**** 50·4**** 14·6**** 15·0**** 0·024·2*  
Prunella modularis3645 27·2**** 86·6****151·5**** 28·2**** 27·6**** 6·8**   
C. flammea 164  0·1 27·6**** 23·6****    0·01 4·0*
Emberiza citrinella1450  6·9** 59·1****   13·5***   

Figure 2. Temporal variation in average clutch size throughout the breeding season in the United Kingdom (•) and New Zealand (○) calculated at 10-day intervals; only intervals in which 10 or more clutches were recorded are plotted. The lines are the fitted relationship between first egg date and clutch size derived from the minimal adequate models for the UK (solid line) and New Zealand (dashed line). First egg dates are measured as days since the winter solstice, taking leap years into account.

Download figure to PowerPoint

Breeding season length was longer in New Zealand than the UK for nine species when measured using the interval index, and for all 11 species when calculated using the MacArthur index (Table 3). Across all species, mean season length was significantly longer in New Zealand (interval index: t10 = 2·48, P = 0·032; MacArthur index: t10 = 6·14, P < 0·0001). Similarly, CSR was smaller in New Zealand than in the UK for 9 of the 11 species (Table 3) and was significantly smaller across all species (t10 = 16·49, P < 0·0001), implying that species exhibit less temporal variation in clutch size throughout the breeding season in New Zealand. Nevertheless, there was no significant relationship between the proportional change in mean clutch size and changes in our measures of seasonality between countries, either with or without taking phylogeny into account (Fig. 3).

Table 3.  Comparison between United Kingdom (UK) and New Zealand (NZ) of three indices of seasonality measured for 11 passerine species. The interval index is the number of days between the 1% and 99% quantiles of first egg dates; the MacArthur index is the number of equally good 14-day periods for nesting (see text). CSR is the ratio of the maximum to minimum clutch size
SpeciesSeason length – interval indexSeason length – MacArthur indexCSR
UKNZ% changeUKNZ% changeUKNZ% change
Fringilla coelebs 67·0 94·0+40·3 5·120·0+293·01·6491·462−11·4
Turdus merula108·0127·5+18·1 8·513·9 +62·61·5681·150−26·6
Sturnus vulgaris 66·0 72·0 +9·1 5·013·6+170·61·8091·431−20·1
Passer domesticus121·0109·5 −9·5 9·618·3 +90·71·5921·301−18·3
Alauda arvensis 89·0109·0+22·5 8·122·6+179·51·4151·149−18·8
Carduelis chloris108·0110·0 +1·910·217·2 +69·21·8861·529−18·9
T. philomelos116·0164·5+41·8 8·015·4 +93·71·6291·295−20·5
C. carduelis105·0117·0+11·4 8·421·3+152·91·8991·546−18·6
Prunella modularis105·0112·0 +6·7 7·419·3+161·91·6331·306−20·0
C. flammea 88·0 96·0 +9·114·523·0 +58·41·7881·506−15·8
Emberiza citrinella 98·0 88·0−10·2 8·416·8 +99·31·4131·191−15·7
Mean  +12·8  +130·2  −18·7
95% confidence interval    1·3 to 24·3    88·6 to 171·7  −21·2 to −16·1

Figure 3. Relationships between seasonality and proportional change in clutch size for 11 passerine species introduced from the UK to New Zealand. Seasonality is measured as (a) CSR, the proportional change in the ratio of maximum clutch size to minimum clutch size during the breeding season, (b) proportional change in breeding season length using the interval index (the interval between the 1% and 99% quantiles of first egg date) and (c) the number of equivalent 14-day periods using the MacArthur index (see text). None of these relationships is significant (Spearman rank correlations, in all cases P > 0·1, controlling for phylogeny does not alter our conclusions).

Download figure to PowerPoint


  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

Consistent with the general pattern that birds in the southern hemisphere lay smaller clutches than their northern counterparts (Moreau 1944; Yom-Tov 1987; Yom-Tov et al. 1994; Martin 1996; Rowley & Russell 1991; van Zyl 1999), 9 out of 11 species in this study laid significantly smaller clutches in New Zealand than the UK. This confirms the trend reported in previous studies comparing clutch sizes of species resident in both countries (Niethammer 1970; MacMillan 1985). New Zealand occurs at slightly lower latitudes than the UK, but this is highly unlikely to explain the reduction in clutch size observed in the multiple regression models as the latter took any significant latitudinal effects into account.

Ashmole's hypothesis interprets variation in clutch size as a consequence of seasonal changes in resource availability. More marked seasonality increases the discrepancy between the size of the breeding population and the breeding environment's carrying capacity, elevating per capita resource availability and promoting larger clutches. New Zealand, being an isolated land mass surrounded by ocean, has a less seasonal climate than the UK. Using data for monthly land surface temperatures on a 10′ latitude/longitude grid (New et al. 2002), the annual range in mean monthly temperature is significantly greater across the UK than New Zealand (UK: mean 11·5 °C, 95% confidence intervals 11·4–11·6 °C; New Zealand mean 10·4 °C, 95% confidence intervals 10·3–10·5 °C), and the UK experiences significantly more ground frosts (mean 109) per year than New Zealand (mean 89). Moreover, lower clutch sizes in New Zealand are associated with longer breeding seasons and less marked temporal variation in clutch size throughout the season, implying that birds experience less seasonal conditions in New Zealand than the UK.

Our results are thus broadly consistent with Ashmole's hypothesis. A shift to less seasonal conditions is matched by an overall decline in clutch size. Nevertheless, if Ashmole's hypothesis was the sole driver of clutch size reductions in NZ, species showing the greatest declines in clutch size should be those experiencing the greatest shift in resource availability from more to less seasonal conditions. We found no relationship between proportional declines in clutch size and proportional changes in the extent to which a species experiences seasonality (CSR, and two measures of breeding season length, Fig. 3). Confirmation of Ashmole's hypothesis would, however, require direct measurement of changes in resource availability, combined with data on seasonal changes in population size and differences in survival rates and population densities between the UK and NZ. Although we lack conclusive data there is some evidence that population densities are higher in New Zealand than the UK, at least for some species (Thomsen 2002; MacLeod et al. 2005). There is also anecdotal evidence that introduced birds in New Zealand live for longer than they do in the UK. For example, the maximum ages reported for banded birds of several introduced passerine species in New Zealand are comparable to ages recorded in the UK, despite a much lower sampling effort in New Zealand (Heather & Robertson 1997).

Other mechanisms may nevertheless also contribute to reduced clutch sizes in New Zealand. One possibility is that our New Zealand study populations have experienced reduced genetic variation as a result of originating from the introduction of few individuals (c. 100–800 birds; Duncan 1997). Subsequent inbreeding depression may have caused reduced clutch sizes (Amos et al. 2001). Evidence for such founder effects is provided by empirical demonstration of a correlation between the size of the founder population and the relative magnitude of increased hatching failure rates in New Zealand (Briskie & MacKintosh 2004). However, proportional changes in failure rates are not significantly related to the reductions in clutch size that we observe (Spearman rank correlation rs = −0·515, P > 0·1, note that controlling for phylogeny using the methods described in Fig. 3 does not alter our conclusions), suggesting that founder effects are not responsible for smaller clutches in New Zealand.

Alternatively, the adult life-span hypothesis (Skutch 1985) suggests that higher survival rates in New Zealand may promote reduced clutch sizes through the trade-off between reproductive effort and survival (Williams 1966a, 1966b; Stearns 1992). This hypothesis is supported by latitudinal gradients in life span that are the inverse of those for clutch size (Moreau 1944; Ghalambor & Martin 2001; Peach, Hanmer & Oatley 2001). Life spans may be higher in New Zealand because of selection for longevity during the lengthy process of transporting founder populations to New Zealand, which imposed high mortality rates (Thomson 1922), or as a consequence of the release from natural enemies, such as parasites. More simply, increased longevity may result from birds experiencing lower winter mortality in less seasonal environments. Such a reduction in winter mortality is a necessary, but not in isolation sufficient, condition for Ashmole's hypothesis.

Finally, several species of small mammal have been introduced to New Zealand, and are now widespread, numerous and active nest predators (King 1990; Sanders & Maloney 2002). Skutch (1949) put forward the hypothesis that gradients in clutch size could reflect gradients in predation pressure, high nest predation rates may favour reduced parental provisioning rates, as this lowers the probability of predators finding nests, thus reducing the optimal clutch size. Skutch's hypothesis has received some experimental support (Ghalambor & Martin 2000; Martin, Scott & Menge 2000b). A study of clutch size variation in North and South American birds (Martin et al. 2000a), however, found that Skutch's hypothesis could explain only intercontinental variation in clutch size and not differences between continents. Comparative data on nest predation rates are scant, but those available suggest that predation may be higher in New Zealand for A. arvensis (Petermann 2003) and E. citrinella exhibits a similar, albeit non-significant, trend (MacLeod et al. 2005).

In addition to documenting geographical variation in clutch size our data illustrate marked geographic variation in its seasonal pattern of variation for 8 of our 11 species; the exceptions being P. domesticus, C. flammea and E. citrinella. In six species an apparent unimodal relationship between clutch size and first egg date in the UK becomes, in New Zealand, a negative linear one. These patterns are compatible with a switch from multibrooded breeding in the UK to raising single broods in New Zealand (Crick et al. 1993). These species also exhibit a reduced rate of decline in clutch size at the end of the breeding season in New Zealand compared with the UK. It has been suggested that the magnitude of decline in clutch size is related to the extent to which early fledged chicks are more likely to survive relative to ones fledged latter in the breeding season (Young 1994); thus late fledged young in New Zealand may have greater survival probabilities than their northern hemisphere equivalents. The seasonal increases in clutch size exhibited by T. merula and A. arvensis in New Zealand are unusual. One possible explanation is that nest predation rates for these species are higher earlier in the breeding season as alternative prey are scarce, but predation rates may subsequently decline as predators switch to other food sources. Although migratory status may influence seasonal patterns in clutch size (Crick et al. 1993; Dhondt et al. 2002), each of our focal species is resident in both countries.

In summary, we demonstrate that within-species clutch sizes are smaller in the southern hemisphere than at equivalent northern latitudes. Our data are consistent with the hypothesis that such variation arises partly as a consequence of reduced seasonality in the southern hemisphere. Additional comparative data on population densities, survival and predation rates are required, however, to fully test this and related hypotheses. Such data may provide further insight into the ‘exciting and challenging’ problem of geographic clutch size variation.


  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

We thank the numerous volunteer nest recorders in United Kingdom and New Zealand who have contributed to two of the longest running nest record schemes in the world, and the Ornithological Society of New Zealand's council for providing access to the New Zealand data. The BTO Nest Record Scheme is funded by a partnership of the BTO and the Joint Nature Conservation Committee. KLE was funded by the Royal Society (UK).


  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References
  • Amos, W., Wilmer, J.W., Fullard, K., Burg, T.M., Croxall, J.P., Bloch, D. & Coulson, T. (2001) The influence of parental relatedness on reproductive success. Proceedings of the Royal Society of London B 268, 20212027.
  • Ashmole, N.P. (1963) The regulation of numbers of tropical oceanic birds. Ibis 103b, 458473.
  • Briskie, J.V. & MacKintosh, M. (2004) Hatching failure increases with severity of population bottlenecks in birds. Proceedings of the National Academy of Sciences of the USA 101, 558561.
  • Butler, S.J. & Gillings, S. (2004) Quantifying the effects of habitat structure on prey detectability and accessibility to farmland birds. Ibis 146 (S2), 123130.
  • Butler, S.J., Whittingham, M.J., Quinn, J.L. & Cresswell, W. (2005) Quantifying the interaction between food density and habitat structure in determining patch selection. Animal Behaviour 69, 337343.
  • Cardillo, M. (2002) The life-history basis of latitudinal diversity gradients: how do species traits vary from the poles to the equator? Journal of Animal Ecology 71, 7988.
  • Christians, J.K., Evanson, M. & Aiken, J.J. (2001) Seasonal decline in clutch size in European starlings: a novel randomisation test to distinguish between the timing and quality hypothesis. Journal of Animal Ecology 70, 10801087.
  • Cooper, C.B., Hochachka, W.M. & Dhondt, A.A. (2005) Latitudinal trends in within-year reoccupation of nest boxes and their implications. Journal of Avian Biology 36, 3139.
  • Cramp, S., Simmons, K.E.L. & Perrins, C.M. (197794) Handbook of the Birds of Europe, the Middle East and North Africa, Volumes 1–9. Oxford University Press, Oxford.
  • Crick, H.Q.P., Gibbons, D.W. & Magrath, R.D. (1993) Seasonal changes in clutch size in British birds. Journal of Animal Ecology 62, 263273.
  • Crick, H.Q.P., Baillie, S.R. & Leech, D.I. (2003) The UK Nest Record Scheme: its value for science and conservation. Bird Study 50, 254270.
  • Dhondt, A.A., Kast, T.L. & Allen, P.E. (2002) Geographical differences in seasonal clutch size variation in multi-brooded bird species. Ibis 144, 646651.
  • Duncan, R.P. (1997) The role of competition and introduction effort in the success of passeriform introductions to New Zealand. American Naturalist 149, 903915.
  • Duncan, R.P., Blackburn, T.M. & Sol, D. (2003) The ecology of bird introductions. Annual Review of Ecology and Systematics 34, 7198.
  • Dunn, P.O., Thusius, K.J., Kimber, K. & Winkler, D.W. (2000) Geographic and ecological variation in clutch size of tree swallows. Auk 117, 211215.
  • Gaston, K.J. & Chown, S.L. (1999) Elevation and climatic tolerance: a test using dung beetles. Oikos 86, 584590.
  • Ghalambor, C.K. & Martin, T.E. (2000) Parental investment strategies in two species of nuthatch vary with stage-specific predation risk and reproductive effort. Animal Behaviour 60, 263267.
  • Ghalambor, C.K. & Martin, T.E. (2001) Fecundity–survival trade-offs and parental risk-taking in birds. Science 292, 494497.
  • Gil-Delgado, J.A., Marco, E., Paredes, M. & Vives-Ferrandiz, C. (2005) Seasonal clutch size variation of multi-brooded bird species: comparisons between breeding season and latitudes. Ibis 147, 206212.
  • Heather, B.D. & Robertson, H.A. (1997) The Field Guide to the Birds of New Zealand. Oxford University Press, Oxford.
  • Hill, R.A., Barrett, L., Gaynor, D., Weingrill, T., Dixon, P., Payne, H. & Henzi, S.P. (2003) Day length, latitude and behavioural (in) flexibility in baboons (Papio cynocephalus ursinus). Behavioral Ecology and Sociobiology 53, 278286.
  • King, C.M. (1990) The Handbook of New Zealand Mammals. Oxford University Press, Auckland.
  • Koenig, W.D. (1984) Geographic variation in clutch size in the northern flicker Colaptes auratus: support for Ashmole's hypothesis. Auk 101, 698706.
  • Lack, D. (1947) The significance of clutch-size. Part I – Intraspecific variations. Ibis 89, 302352.
  • Lack, D. (1948) The significance of clutch-size. Part III – Some interspecific comparisons. Ibis 90, 2545.
  • Lack, D. & Moreau, R.E. (1965) Clutch-size in tropical passerine birds of forest and savanna. Oiseau (Review of French Ornithology) S35, 7689.
  • MacArthur, R.H. (1964) Environmental factors affecting bird species diversity. American Naturalist 98, 387397.
  • MacLeod, C.J., Duncan, R.P., Parish, D.M.B., Wratten, S.D. & Hubbard, S.F. (2005) Can increased niche opportunities and release from enemies explain the success of introduced Yellowhammer populations in New Zealand? Ibis 147, 598607.
  • MacMillan, B.W.H. (1985) Breeding of greenfinches in Hawke's Bay, New Zealand. Notornis 32, 8593.
  • Martin, T.E. (1996) Life history evolution in tropical and south temperate birds, what do we really know? Journal of Avian Biology 27, 263272.
  • Martin, T.E., Martin, P.R., Olson, C.R., Heidinger, B.J. & Fontaine, J.J. (2000a) Parental care and clutch sizes in North and South American birds. Science 287, 14821485.
  • Martin, T.E., Scott, J. & Menge, C. (2000b) Nest predation increases with parental activity: separating nest site and parental activity effects. Proceedings of the Royal Society of London B 267, 22872293.
  • McEwan, W.M. (1987) Ecological Regions and Districts of New Zealand, 4 Parts. Department of Conservation, Wellington.
  • Moreau, R.E. (1944) Clutch-size, a comparative study, with special reference to African birds. Ibis 86, 286347.
  • New, M., Lister, D., Hulme, M. & Makin, I. (2002) A high-resolution data set of surface climate over global land areas. Climate Research 21, 125.
  • Niethammer, G. (1970) Clutch sizes of introduced European passeriformes in New Zealand. Notornis 17, 214222.
  • Peach, W.J., Hanmer, D.B. & Oatley, T.B. (2001) Do southern African songbirds live longer than their European counterparts? Oikos 93, 235249.
  • Petermann, J. (2003) Breeding success and summer foraging behaviour of skylarks Alauda arvensis on the Canterbury Plains, New Zealand. Diplomarbeit, University of Bremen, Germany/Lincoln University, New Zealand.
  • Purvis, A. & Rambaut, A. (1995) Comparative analysis by independent contrasts (CAIC): an Apple Macintosh application for analysing comparative data. Computer Applications in the Biosciences 11, 247251.
  • Ricklefs, R.E. (1980) Geographical variation in clutch size among passerine birds, Ashmole's hypothesis. Auk 97, 3849.
  • Ricklefs, R.E. (2000) Density dependence, evolutionary optimization, and the diversification of avian life histories. Condor 102, 922.
  • Robertson, H.A. (1986) OSNZ nest record scheme instructions. Notornis 33, 5357.
  • Rowley, I. & Russell, E. (1991) Demography of passerines in the temperate southern hemisphere. Bird Population Studies (eds C.M.Perrins, J.-D.Lebreton & M.Hirons), pp. 2224. Oxford University Press, Oxford.
  • Sanders, M.D. & Maloney, R.F. (2002) Causes of mortality at nests of ground-nesting birds in the Upper Waitaki Basin, South Island, New Zealand: a 5-year video study. Biological Conservation 106, 225236.
  • Sanz, J.J., Tinbergen, J.M., Moreno, J., Orell, M. & Verhulst, S. (2000) Latitudinal variation in parental energy expenditure during brood rearing in the great tit. Oecologia 122, 149154.
  • Sibley, C.G. & Ahlquist, J.E. (1990) Phylogeny and Classification of Birds: a Study in Molecular Evolution. Yale University Press, New Haven, CT.
  • Skutch, A.F. (1949) Do tropical birds rear as many young as they can nourish? Ibis 91, 430455.
  • Skutch, A.F. (1985) Clutch size, nesting success, and predation on nests of Neotropical birds, reviewed. Ornithological Monographs 36, 575594.
  • Stearns, S.C. (1992) The Evolution of Life Histories. Oxford University Press, Oxford.
  • Thomsen, S. (2002) The Ecology of the Skylark Alauda arvensis on the Canterbury Plains, New Zealand. PhD Thesis, Lincoln University, New Zealand.
  • Thomson, G.M. (1922) The Naturalisation of Plants and Animals in New Zealand. Cambridge University Press, Cambridge.
  • Whittingham, M.J. & Evans, K.L. (2004) The effects of habitat structure on predation risk of birds in agricultural landscapes. Ibis 146 (S2), 210220.
  • Williams, G.C. (1966a) Adaptation and Natural Selection. Princeton University Press, Princeton, NJ.
  • Williams, G.C. (1966b) Natural selection, the costs of reproduction, and a refinement of Lack's principle. American Naturalist 100, 687690.
  • Yom-Tov, Y. (1987) The reproductive rate of Australian passerines. Australian Wildlife Research 14, 319330.
  • Yom-Tov, Y., Christie, M.I. & Iglesias, G.J. (1994) Clutch size in passerines of southern South America. Condor 96, 170177.
  • Young, B.E. (1994) Geographic and seasonal patterns of clutch-size variation in house wrens. Auk 111, 545555.
  • Van Zyl, A.V. (1999) Breeding biology of the Common Kestrel in Southern Africa 32 degrees S. compared to studies in Europe 53 degrees N. Ostrich 70, 127132.