Most hummingbirds migrate in two broad waves tied to day length and food supply: northbound movement runs roughly February through May across North America, and southbound movement runs roughly July through early November, with ruby-throated hummingbirds leaving the eastern United States mostly between late August and mid-October. Peak timing shifts by latitude, species, and sex, and the trigger is not cold itself but the shrinking daylight and dwindling nectar that precede it.
Understanding when hummingbirds migrate means abandoning the idea of a single calendar date. Migration is a rolling series of local departures and arrivals, staggered by latitude, age, body condition, and the flowering schedule of the landscapes the birds cross. A hummingbird in southern Louisiana may head south weeks after one in Ontario has already gone, and a young bird may leave days after adults of the same species have cleared out.
The physiological trigger: fat, daylight, and torpor
A hummingbird does not decide to migrate because it feels chilly. The proximate trigger is photoperiod, the changing ratio of daylight to darkness, which shifts the endocrine system toward migratory restlessness, called zugunruhe, and toward hyperphagia, a dramatic increase in feeding. In late summer, ruby-throated hummingbirds roughly double their body mass, storing fat as the primary fuel for overwater crossings. A bird that weighs about 3 grams in June may weigh close to 6 grams in September. That fat reserve is what carries it across the Gulf of Mexico, a nonstop flight of roughly 500 to 600 miles that can take 18 to 22 hours.
Fueling this kind of flight requires extraordinary metabolic output. Hummingbirds sustain some of the highest mass-specific metabolic rates of any vertebrate, and they must do so while their wings beat tens of times per second. To survive cold nights and fasting periods, they enter torpor, dropping body temperature and metabolic rate, then rewarming at dawn. A migrating bird that arrives at dusk with depleted reserves can survive an unexpectedly cold night only because torpor buys it time.
The timing logic is simple once you see the fuel constraint. Day length is a stable astronomical signal that tells the bird weeks in advance what the food supply will soon do. Fat is the currency that converts that signal into a flight. One plain-language summary: the bird reads the calendar in the sky, then fills its tank before the gas station closes.
Quantified context: dates, distances, and body masses
Across temperate North America, the broad timing bands are consistent. Northbound arrivals in the southern states typically begin in February and March; the upper Midwest and Northeast see birds mostly from mid-April through mid-May. Southbound, the Gulf Coast sees adult male ruby-throats moving through from late July, adult females in August, and juveniles lingering into October. Most ruby-throated hummingbirds have left the eastern United States by mid-October, though scattered individuals persist at feeders through November and, rarely, into winter.
The distances are real but not uniform. Ruby-throated hummingbirds breeding in Canada may fly 2,000 to 3,000 miles to wintering grounds in Mexico and Central America. Rufous hummingbirds, by contrast, breed unusually far north, reaching Alaska and western Canada, and migrate to winter in Mexico, with some individuals wintering along the Gulf Coast. That north-south span of roughly 3,000 miles is one of the longest migratory journeys relative to body size among birds. Only a handful of species, including the ruby-throated, make the Gulf crossing nonstop; many western species follow routes that keep land within reach.
Body mass figures anchor all of this. Most hummingbirds weigh between 2.5 and 6 grams, and the fat loads that support a Gulf crossing represent a near-doubling of that mass. Those numbers explain why departure timing is so tightly coupled to food availability along the route, and why a cold snap or a failed bloom can be catastrophic for newly arrived birds.
Edge cases and counterintuitive details
The cleanest counterintuitive point is that not every hummingbird migrates. Several species, including Anna's and Allen's hummingbirds, are largely resident where winters are mild, and a few individuals overwinter in places like the Pacific Northwest and the Gulf states. A feeder in a mild-winter yard may host the same bird all year. This blurs the tidy expectation that all hummingbirds leave, and it matters to anyone deciding when to take feeders down.
A second edge case is differential timing by sex and age. Adult males typically migrate first, followed by adult females, then juveniles. This staggered departure is measurable in feeder traffic: a quiet week in early August may mean the males have gone, not that birds have vanished. Similarly, young birds may stay weeks longer than adults, and some late nesters may delay departure to finish fledging. A third edge case is that migration timing can shift with weather, particularly in the western United States, where the timing and volume of monsoon rains influence regional blooms and, in turn, local movement.
A professional caveat: some apparent "departures" are actually survival failures, not migration. Birds that fail to build fat reserves do not simply migrate late; they may not migrate at all, and observers rarely see that outcome. Counts at feeders underrepresent mortality and overrepresent successful, well-fueled individuals.
Misconceptions and red flags
The most persistent misconception is that hummingbirds migrate because feeders fail or because temperatures drop below a specific threshold. Feeders are not a cause of migration, and leaving feeders up will not delay departure. Keeping feeders up into autumn is encouraged precisely because late or weather-delayed migrants may still need food. If a hummingbird does not move south on schedule, the reason is usually body condition or an unusual weather pattern, not a feeder.
A second misconception is a single national departure deadline. The internet often repeats one date, but timing varies by hundreds of miles and several weeks. Extending a Gulf Coast departure date to the northern tier is simply wrong.
Watch for these red flags:
Treating a late-summer lull as evidence that all birds have gone. Assuming a bird at a feeder in November must be lost or sick rather than a normal late or overwintering individual. Rigidly taking feeders down on a fixed date regardless of latitude and local weather. Assuming every hummingbird species makes the same long overwater flight. Interpreting regional declines in feeder visits as population declines without other evidence.
Comparison of migration strategies
| Species group | Typical wintering location | Northbound timing | Southbound timing | Overwater crossing |
|---|---|---|---|---|
| Ruby-throated (eastern US) | Mexico, Central America | Feb–May | Late Jul–Oct | Yes, Gulf of Mexico |
| Rufous (western) | Mexico, Gulf Coast | Mar–May | Jul–Sep | Sometimes |
| Anna's (Pacific) | Resident in mild areas | Year-round | Year-round | No |
| Broad-tailed (Rockies) | Mexico | Apr–May | Aug–Sep | No |
| Black-chinned (western) | Mexico | Apr–May | Jul–Sep | Sometimes |
The table makes the practical point: a single "migration window" is an artifact of averaging. Each species and region has its own schedule, and the overwater column determines how tightly that schedule is coupled to body condition.
Timeline of a single adult male ruby-throated hummingbird
| Period | Condition | Behavior |
|---|---|---|
| June | Breeding on territory | Defends flowers, low body mass |
| July | Late breeding, post-breeding | Begins hyperphagia |
| Early Aug | Fat loading | Mass rises toward roughly 5–6 g |
| Mid-Aug | Departure window | Southbound movement begins |
| Late Aug–Sep | Gulf staging | Fueling along coastal stopover |
| Sep–Oct | Overwater crossing or coastal route | 18–22 h nonstop if crossing |
| Oct–Nov | Arrival on wintering grounds | Territory establishment |
Practical actions
The timing question matters most to people managing feeders and habitat. The evidence supports a few concrete practices.
Keep feeders up for at least two weeks after your last observed hummingbird, and preferably into late November in most of the United States. The extra window catches late juveniles, weather-delayed migrants, and overwintering individuals. Clean and refill feeders on a schedule that matches consumption, and use a simple sugar solution without additives such as red dye or concentrated commercial mixes of uncertain composition.
If you want to track local timing, record first and last sighting dates, note whether the bird is male or female, and report to a regional monitoring project. That observer data is how migration timing gets refined, and it is the cleanest way to reconcile the generic date ranges in this article with your own yard. For late birds, keep a feeder accessible but do not assume a stranded bird needs rescue; contact a licensed wildlife rehabilitator only if the bird appears injured or unable to fly.
One threshold worth holding onto: a ruby-throated hummingbird that stays past mid-October in the northern United States is unusual but not impossible, and it should be reported rather than ignored. Two to four weeks after that is when most records stop.
Limitations and what the numbers cannot tell you
Migration dates are population-level summaries, not individual guarantees. The bands cited here come from broad regional patterns, and the exact departure of any one bird depends on its body condition, the flowering of plants along its route, and weather that models cannot fully predict. Timing estimates also vary by survey method: feeder observations, banding data, and radar detect different subsets of birds, so they rarely agree perfectly. The claim that feeders do not delay departure is well supported, but the claim that they are never beneficial is not; the responsible reading is that feeders are neutral to helpful, never harmful in themselves.
Date ranges should be treated as typical, not absolute. Climate variability is shifting some species' schedules, and a single year's observations cannot establish a trend. When a late bird appears, the more careful conclusion is "this is within the range of recorded variability," not "this is proof of a new normal."