The Evening Grosbeak was once rare east of the Rockies, but historically its range has shifted, perhaps in response to changing food supplies. Mike Hopiak/CLO

When I was in grade school, every winter my parents' feeders in Alberta, Canada, were swarmed by Evening Grosbeaks consuming sunflower seeds like Hoover vacuum cleaners on wings. By the time I was attending university, the golden avian hordes were a thing of the past, and it was rare even to hear a lone grosbeak flying high overhead. It wasn't until I started working at the Lab of Ornithology that I appreciated how typical my observations were. Comments by coworkers and observations by FeederWatchers suggested that there were large areas where Evening Grosbeaks had become less frequent winter visitors. And by using data from thousands of FeederWatchers, I was finally able to determine the extent of these declines.

The data from long-term participants of Project FeederWatch (PFW) revealed that, from 1992 through 2001, the likelihood of seeing Evening Grosbeaks indeed declined in some regions, including most parts of Canada and the midwestern and southeastern United States (Figure 1). However, in other areas, such as the northern Appalachian and western Great Lakes, Evening Grosbeaks became more frequent visitors at feeders of long-term participants.

Would the same patterns have been found using all of the PFW data, from both short- and long-term participants? No. For example, Figure 2 shows data from FeederWatchers regardless of how many years they submitted data. These data do show some biologically real patterns, such as year-to-year oscillations reflecting periodic irruptions of Evening Grosbeaks. But compare the long-term trends of, say, New England in Figure 2 with the more reliable patterns generated from long-term observers in Figure 1. The graph based on all data suggests a stable or decreasing prevalence, but the data from long-term participants indicates that, on the contrary, Evening Grosbeaks in New England have increased in prevalence.

Figure 1. In some regions of Canada and the United States, the likelihood of seeing grosbeaks at feeders has declined since 1992, while in other areas it has increased. The colors on the map denote the average change in the probability that a long-term (five or more years) FeederWatcher saw at least one Evening Grosbeak between December and February each year, 1992-2001. "No data" indicates no long-term data available or no reports of Evening Grosbeaks.

This difference occurs not because data reported by short-term participants are wrong, but simply because the trends in Figure 2 are a combination of fluctuations in grosbeak abundance (such as the irruptions) and changes in the numbers and locations of participants through time. For example, a recent increase in the number of short-term FeederWatchers in areas with only rare reports of grosbeaks could generate declining trends caused by changes in the distribution of participants rather than of grosbeaks.

Because we used data from long-term participants to generate Figure 1, we can be confident that the results reflect real biological patterns. These patterns of changing distributions are in keeping with a historic tendency for Evening Grosbeaks ranges to shift. In the mid-1800s, Evening Grosbeaks were reportedly rare or uncommon east of the Rockies; in Quebec nesting wasn't documented until the 1940s. Some researchers speculate that as planting box elder trees became more popular, the new source of seeds enticed the birds eastward. Another supposition is that spruce budworm outbreaks in the East provided a food supply that coaxed grosbeaks into new areas.

Figure 2. Using all available data on Evening Grosbeaks, even from participants who submitted data in just one year, oscillations in abundance from year to year are apparent in some regions, such as New England and the North Atlantic, reflecting periodic winter irruptions. However, the apparent long-term trends often differ from those based on more robust analyses using only long-term participants' data (see Figure 1).

The recent changes in winter distribution may result from changing migratory patterns. Changes in the breeding distribution may be another factor. However, when the winter data are compared with breeding data from North American Breeding Bird Survey maps, increases and declines in summer do not map closely onto the winter declines.

Based on the data from long-term FeederWatchers, I now know that the decreasing frequency of Evening Grosbeaks I observed at my parents' feeders was part of a regional trend of decreasing winter prevalence over the last decade or more. However, it's reassuring that, despite these regional changes, overall Evening Grosbeaks do not appear to be vanishing from the continent as a whole.