Cornell Lab of Ornithology

bslogo.gif (22710 bytes)

The Birdhouse Network
Become A Member

Clutch-size Variation in Eastern Bluebirds
BY André A. Dhondt Tracey L. Kast, and Paul E. Allen

Please cite this Page as:
Dhondt, A. A. ,   Kast, T. L. and Allen, P. E. 2000.  Clutch-size Variation in Eastern Bluebirds. Birdscope, Volume 14, Number 2:  3-5.

Birdhouse monitors document previously unknown results

When we began the Cornell Nest Box Network (CNBN) three years ago, we were quite ambitious: we wanted to study how clutch size (the number of eggs females lay in one nest attempt) varied geographically, the effects of blowflies on reproduction (see the Autumn 1999 issue of Birdscope), and possible effects of calcium deficiency on egg quality (see the Autumn 1999 and the Winter 1998 issues of Birdscope). These studies simply required observations around nest boxes (clutch size, blowflies) or simple experiments (calcium) by project members. On top of that, we also began studies of dispersal by training large numbers of citizen scientists in the art of bird banding.

Although the number of participants in CNBN was not as large as we had hoped, we did obtain a substantial amount of data on the reproductive biology of cavity-nesting birds. This article summarizes data pertaining to factors that influence clutch size variation in Eastern Bluebirds (Sialia sialis).

Our main questions were, would clutch size in Eastern Bluebirds increase from south to north and from east to west? Although such a pattern has been found in some bird species, literature on whether such a pattern generally exists in most bird species has been controversial. Regarding the Eastern Bluebird itself, an earlier analysis of the Lab of Ornithology’s Nest Record Cards published in Living Bird in 1970 could not document significant clutch size variations with latitude (D. B. Peakall. 1970. “The Eastern Bluebird: Its Breeding Season, Clutch Size, and Nesting Success.” Living Bird 9, 239-256.).

One reason there is still so much controversy about geographic variation of clutch size is that many other factors can influence clutch size, and often not all factors are taken into account when analyzing data. For example, in many bird species older females, or birds that breed early, at lower altitudes or in larger nest boxes tend to lay larger clutches than younger females or those that lay later in the season, at higher elevations or in smaller cavities. Furthermore, birds tend to lay more eggs in some years than in others. To carry out a meaningful analysis, researchers need a large data set, spread over the entire breeding season. That way, even though we don’t know the age of the female or the quality of her territory, we can partition out possible effects of egg-laying date, year, altitude, or nest-box volume.

In 1997 and 1998, CNBN participants reported data on 2,234 nesting attempts of Eastern Bluebirds as far south as Florida and Texas and as far north as North Dakota and the Canadian provinces of Quebec and Ontario. To be certain that the clutches we analyzed were complete clutches, we included only data from nests in which young had been observed. That left us with 611 clutches for which we had information on all of the variables we wanted to study. Because the first analysis showed no effect for nest box volume or altitude, we could then add 334 more observations. The distribution of reported latitudes (29.88-48.60º N) and longitudes (103.34-70.03º W) spanned the Eastern Bluebird’s range in North America, although the majority of observations came from the northeastern United States. The results described below are based on the 945 observations for which we had information on clutch size, egg-laying date, latitude, and longitude. The mean clutch size was 4.49 eggs, with extremes of 2 and 7 eggs. We calculated whether latitude, longitude, egg-laying date, and year in which the clutch was laid had significant effects on clutch size.

The results? As explained in Figure 1, we needed to use complex statistical techniques and generate rather complicated results to answer a simple question: Does clutch size vary with geographic location? The quick answer is yes. Clutch size increases from south to north and from east to west.

Figure 1
smfig1eabb.gif (3553 bytes)
Figure 1. These four graphs illustrate different models for describing latitudinal variation in clutch size. In all cases, clutch sizes vary across the season. The red lines represent clutch sizes farther south, and the black lines represent clutch sizes farther north. In these hypothetical examples, we do not consider possible effects of year or of longitude.

Panel a illustrates the simplest model. Here, clutch sizes become smaller at a constant rate as the season progresses, and this linear change is represented by a straight line. The rate of change of clutch size with date is the same at all latitudes, and therefore the lines are parallel. Statistically, we say that clutch size is a function of date and latitude.

Panel b illustrates the case where clutch sizes are largest in the middle of the season, but where the shape of the curve describing that relationship does not vary with latitude. The lines representing the relationships at different latitudes have the same shape and are parallel. Mathematically, this curve is generated by adding the term date squared to the model. Here, clutch size is a function of date, date squared, and latitude.

In Panel c, as in Panel a, clutch size decreases linearly with data, but the rate of change of clutch size with date varies with latitude. The slopes of the lines describing clutch size as a function of date vary with latitude, and the lines, therefore, are no longer parallel to each other. Mathematically, this model shows that clutch size is a function of date, latitude, and the interaction between date and latitude.

Panel d combines the effects illustrated in panels b and c. The model describing clutch size shows effects of date, of date squared (so that the curves are not linear), and of latitude; it also includes interaction terms that result in different curve shapes at different latitudes (interactions between date and latitude and between date-squared and latitude). The model describing the real data in Figure 2 also includes year, longitude, and year-by-latitude terms.

Longitudinal variations were the same in both years and across all latitudes. We can therefore say that per degree of longitude the average clutch size of Eastern Bluebirds increases by 0.014 eggs. Over the Eastern Bluebird’s geographic range, this is not insignificant. All else being equal, a participant in Maine at 70° W would report a clutch size that was smaller by 0.46 eggs than a participant in North Dakota at 103° W.

The effects of latitude are more complex because clutch size also varies with date. But how clutch sizes vary with date depends on the latitude at which we make our observations. In the southern part of the range, the early clutches are relatively small, clutch sizes reach a peak in the middle of the season, and clutches laid late in the season are small again. In the northern part of their range, however, clutch sizes decrease gradually with the laying date. The earliest clutches are, on average, the largest, and as the season progresses, females lay smaller and smaller clutches. This is illustrated in Figure 2 for birds breeding at 80° W. Because of differences in the shape of the clutch size/date relationship, differences in clutch size among birds breeding at different latitudes were smallest in May but much larger early and late in the season. This raises questions about what environmental factors cause geographical differences in clutch size. With each new nesting season and more data, we will explore these questions.

Figure 2
fig2eabb.gif (6337 bytes)
Figure 2. Eastern Bluebird clutch size as predicted from the model using the 945 clutches reported by CNBN participants. The model represents values for the year 1997 at 80° W and at 30° N and 45° N. Clutch sizes farther south are small early in the season, increase to a midseason peak, and become smaller again late in the season. Clutches farther north decline throughout the season from an initial high. The graph illustrates the significant latitude-by-lay-date and latitude-by-lay-date-squared interaction terms (see Figure 1). The Julian lay date expresses date as days since January 1.

Finally, we found that clutches laid in 1997 were different from those laid in 1998. Thus, at a given location, early clutches were larger in 1998 than in 1997, whereas for clutches laid late in the season, the opposite was true (see Figure 3).

Once more these results illustrate the power of citizen science. By coordinating the efforts of hundreds of careful observers, we have documented a previously unknown result. If we continue to collect these kinds of data, we should be able to determine, for example, to what extent global warming has variable effects for different species and for a single species in different parts of its range (see cover story). We can also ponder why our study revealed a pattern in Eastern Bluebirds that had remained hidden until our project took flight. We believe there are several reasons for this. One is that in previous analyses, data from all years were lumped together, and we found from this study that clutches varied between years. We now have powerful statistical tools at our disposal that permit the kind of complicated analysis that was needed for this report-tools unavailable just five years ago. Having data collected by participants all across the country enabled us to compare differences in clutch size across the entire range of the Eastern Bluebird.

Figure 3

fig3eabb.gif (6549 bytes)

Figure 3. The Eastern Bluebird predicted clutch size at 100 W, 30 N in 1997 and 1998. This graph illustrates a year-by-lay-date interaction. Clutches laid ealy in the season were larger in 1998 than in 1997,  whereas clutches laid late in the season were larger in 1997 than in 1998. The Julian lay date expresses date as days since January 1.

These initial findings are exciting and unprecedented, and we thank all of our participants for monitoring their nest boxes and providing the data that made this analysis possible. We hope everyone who has nest boxes will sign up for The Birdhouse Network (TBN), our new name for the Cornell Nest Box Network. With your participation and valuable data, TBN will continue to examine environmental factors affecting clutch size. If you are not a member of The Birdhouse Network but would like to join, please visit the web site at, send e-mail to us at, or call (800) 843-BIRD.