Your FeederWatch count site should be an area that is convenient to observe, such as a backyard outside a kitchen or dining room window. Select an area as large as you can consistently observe from week to week. Your count site should include feeders, a water feature, and/or plantings that you maintain for birds. We prefer that you select a count site that you can observe from one vantage point, but multiple vantage points are acceptable so long as you never add counts together (if you see a male cardinal out your front window and a female cardinal out your back window, you should only report one cardinal because that’s the most you saw from one vantage point at one time). Choose obvious boundaries, such as the border of your yard or an area within a courtyard. Most count sites are roughly the size of two tennis courts, but can be as small as a single feeder. Once you have chosen your site, count birds at this same site all season.
Visit the Feeding Birds section for more information about how to create a count site that will be attractive to birds.
The FeederWatch season always begins the second Saturday in November and runs for 21 weeks, ending on a Friday. The 2020–21 FeederWatch season begins on November 14 and ends on April 9. The last day to start a two-day count is April 8.
FeederWatch count days are two consecutive days when you count the birds at your feeders. Count days always come in pairs. Pick days that will maximize the time you have to count birds. Schedule your count days in advance, if possible. Do not change your count days just because you see remarkable numbers or kinds of birds. Doing so would bias your data. If you are unable to count during a particular week or count period, that’s okay. Your data are valuable even if you were only able to count on a few occasions.
Schedule each pair of count days at least 5 days apart (leave five days when you don’t count birds between each count). You may decide to count every Tuesday and Wednesday, for example. Counting once a week means that you can submit up to 21 bird counts–the number of weeks in the FeederWatch season. But it is fine to miss weeks. The more often you count, the more valuable your data becomes, but there is no minimum required number of counts.
Watch your feeders for as long as you can during your count days. However, you don’t need to get up early or watch continuously. Some people can only watch before and/or after work, for example. If you cannot count during both of your count days, try to count as long as possible on your one available day.
Be sure to keep track of how much time you spend observing your count site. For example, if you are in the kitchen working for an hour, and you can see the feeders out your kitchen window, you would only record the portion of that hour when you were actually looking at the feeders.
During each two-day count, you should keep a tally of the birds that visit your site. You can keep track using the FeederWatch mobile app or a Tally Sheet.
To keep track of your counts on paper, use the links below to download one of our Tally Sheets. You can open the plain text version in a word processing document and type in the species that usually visit your yard before printing. Or you can print a PDF version and write in your most common species before photocopying it.
Correctly identifying the birds at your feeder is critical to the quality of the data you submit and to the success of Project FeederWatch. Since relatively few species of birds visit most feeders, these species can become very familiar to you with a little practice and careful observation. We encourage you to acquaint yourself with the birds in your area by studying the Common Feeder Birds Poster included in your first research kit. You can download a mini version of the poster for free.
If you cannot identify a bird, sketch or photograph the mystery bird and pay attention to its field marks before consulting a field guide. If you see the bird during a FeederWatch count, include notes about when you saw the bird and how many you saw so that you can add the observation to your count if you are able to identify the species later.
To identify a mystery bird, consult a current field guide, such as the Cornell Lab’s All About Birds online guide. For more information about identifying birds, visit Identifying Birds in the Learn section of this web site. For help with similar looking birds, such as finches, woodpeckers or accipiters, visit the Tricky Bird ID page. You can also use the Cornell Lab’s Merlin app for help identifying birds.
To ensure that FeederWatch data can be used for scientific research, every FeederWatcher must count birds in the exact same way. Here’s how to conduct your two-day count:
FeederWatch participants often stop counting their birds because they believe that their counts are not important. Typically they are seeing the same birds every week, or they are seeing very few or no birds. While we are all thrilled by unusual sightings and high counts, it’s the everyday observations of common birds that are so important for monitoring bird populations. The only way scientists know when birds are missing is if the people seeing no or very few birds tell us. Learn more about why every count matters.
Remember if no birds visit your feeders, this information is important. If you see no birds, please select “I watched my feeders, but no birds were present” at the top of your count list when you submit your count.
For further information and tips on count procedures, please review Tricky Counts and Special Cases.
Although we are able to access extensive weather information and match it to your counts, snow depth is highly variable from one site to the next and can be important for interpreting patterns of feeder use. Similarly, information on the amount of time that you spent watching your feeders during each count is important to our interpretation of your data.
Record the average depth of snow at your FeederWatch count site over the two-day count. If your count site contained snow drifts of various depths, or if you had snow on one of your count days but not the other, average the depth of snow cover over the two-day count.
FeederWatch participants are invited to report two types of interactive behaviors they might observe between birds at their feeders: displacement and predation (defined below). If you observe either of the behaviors, note on your tally sheet the species of the bird attempting to displace or depredate another bird and the species of the bird that was targeted. They could be the same species or different species. Also record whether or not the behavior was successful (e.g. did the Blue Jay successfully cause the chickadee to leave? Did the Cooper’s Hawk fail to catch the Mourning Dove?).
Displacement is when one bird (the “source”) tries to take over a resource (usually food, but sometimes a perch) occupied by another bird (the “target”). A displacement event is successful if the source bird dislodges the target bird from a perch or feeder. The source bird needs to be purposefully attempting to take the perch of the target bird, rather than landing on a spot as the target bird was about to leave on its own accord. Displacement behavior does not include when a bird flies away to escape a predator or when a group of birds mobs another bird. Sometimes large birds, such as Blue Jays or Red-bellied Woodpeckers, can arrive suddenly at a feeder and cause other birds to scatter. Or sometimes a flock of birds, such as Bushtits, can arrive and cause other birds to leave. Such instances are difficult to interpret so we ask that you only report clear examples of one individual attempting to displace another individual.
Predation events are when one bird (the “source”) attempts to capture or kill another bird (the “target”). Predation events often happen quickly and can be difficult to see. If you are confident that the source species killed the target species, then report the predation event as successful (it is successful from the predator’s perspective). When an event is unsuccessful, such as when a Cooper’s Hawk flies at a bird feeder and comes up empty, it may not be possible to identify the target species. Only record observations where you are certain about the identity of both species.
A note about terms: the verb predate means to come before. Therefore, we use the verb depredate to describe the behavior seen in a predation event.
It can be very difficult to count large or even small flocks of birds, especially when they keep moving around. To estimate the number of birds in a flock, use the “blocking” method. First count the birds in an imaginary block of typical density. Keep the block small, to include only 10 to 25 birds. Then visually superimpose the block onto the entire flock and estimate how many times it fits. Finally, multiply this number by the number of birds in the original block. To get the best estimate, repeat this procedure at various times throughout the day and average your results.
When large, mixed-species flocks appear in your yard, keeping track of the kinds and numbers of birds can be difficult. First, estimate the total flock size using the method above. For example, estimate a flock at 80 birds. Now, take several “samples” of those 80 birds, such as small groups that are easily visible under the feeder, and estimate the proportion of each species in each group. For example, a group of 10 birds might include 5 Dark-eyed Juncos, 4 American Tree Sparrows, and 1 White-throated Sparrow. If that group seems representative of the entire flock, apply your calculated proportion to the total of 80 birds, and estimate the total flock at 40 juncos (50% of flock), 32 Tree Sparrows (40% of flock), and 8 White-throated Sparrows (10% of flock).
Some species are “sexually dimorphic,” that is, the male and female look different. An example is the Northern Cardinal. Some days, the male and female both may visit your feeder, but they never appear at the same time. Obviously you have two different cardinals in your yard. Still, you can only count them as two individuals if you see them together at the same time. Why? Because for FeederWatch data to be scientifically valid, participants must follow the exact same counting procedure for all species.
These species are difficult to tell apart. Even knowing their songs and calls won’t help every time because they can learn each other’s vocalizations. If you live near the area where the ranges of these species overlap (see map at right) and are uncertain which species is at your feeders, please record your chickadees as “Black-capped Chickadee/Carolina Chickadee.”
When a mixed-species flock in your count site is joined by a species that is not typically seen at feeders, you may count the “tag-along bird” even if it doesn’t actually visit your feeder. (The bird was indirectly attracted to your feeder site.)
To help us learn more about feeder birds, we need your data–even if you made just one count! If you signed up for Project FeederWatch through our online store, or if you signed up by mail or phone and received your instructional kit in the mail with your ID number, you are ready to plot your count site on our map and start entering counts! Go to the Your Data section of our website and follow the onscreen instructions.
If you have not yet signed up, join now.
Have you had one of those banner FeederWatching days when a flock of Cedar Waxwings descends upon your yard? More than 100 waxwings are dripping from the trees and shrubs and frolicking in your birdbath. Or perhaps you found a lost bunting at your feeders in the middle of the winter when the bird should be in Central America? These are indeed exciting events and you may be eager to enter your count into the FeederWatch data entry system. As you type in your report, you receive a message on the screen that says, “This is a high count for this species. Please confirm.” Did you do something wrong? Probably not, but you have been introduced to the FeederWatch system designed to check and “flag” potential errors.
We have all made mistakes, either in identifying species or in hitting the wrong key when entering our counts. The data flagging system is designed to catch errors before they are permanently added to the FeederWatch data base.
A computer (called the review robot) automatically screens all counts. It compares your counts to a series of allowable species/maximum count combinations called “filters.” The filters are based on the counts submitted by past FeederWatchers in your area. Although a species or a high count may not be unusual in your yard, it may be unusual when compared with the reports of others in your area. Congratulations! You have experienced something that others in your area have not. When the review robot finds a count that exceeds the maximum set by the filters, it generates a confirmation message and “flags” the count for review.
Bird ranges are dynamic and the filters are designed to accommodate changes in bird distributions. For instance, as Eurasian Collared-Doves rapidly expanded their range across the country, FeederWatch staff were able to revise checklists to minimize the number of participants who were asked to confirm their Eurasian Collared-Dove reports.
Unusual experiences certainly occur at feeders all the time–these rare events keep many of us watching in anticipation of what may happen next. The flagging system is designed to help us recognize when a report is unusual for an area and to help ensure the accuracy of the FeederWatch data so that we may learn more about the birds that we all enjoy.
If the filter considers your count to be unusually high, you will be asked to confirm the entry. Did you intend to type 100, or should it have been 10? If the count is correct, simply click the “Confirm” box, and the data will go to the FeederWatch database at the Cornell Lab of Ornithology. There is no need to do anything further, such as sending photos or correspondence explaining your count, unless a FeederWatch staff member contacts you.
If the filter considers the species you are reporting to be rare for your area, you will be asked to submit a photo or description of your sighting.
After a record is flagged and confirmed, the record goes through a review system managed by FeederWatch staff at the Cornell Lab and Birds Canada. The experts often immediately recognize these flagged records as valid sightings and clear the flags. For the majority of flagged reports, that’s the end of the story. However, if you report a species or count that is rare for your area, you may receive a message from FeederWatch staff asking you to provide more details about your report. A photo is necessary to confirm extremely rare reports. Birds often show up in unusual places, and we are as excited as the participants about these rare sightings.