Open Field Protocol

Nectar Spoilage Timeline by Temperature

The standard advice is to change hummingbird nectar every 2 days above 85°F (29°C). This protocol measures how many hours a standard 1 to 4 batch stays visibly unspoiled at the temperature beside your own feeder, in sun and in shade.

The question this answers#

Every hummingbird page here, including the nectar recipe, says change nectar every 2 days above about 85°F (29°C) and every 3 to 4 days in mild weather. That one sentence serves a reader in Phoenix at 108°F (42°C) and a reader in Nova Scotia at 61°F (16°C).

This protocol replaces the rule with a measurement: the hours between filling a clean feeder with fresh 1 to 4 nectar and the first check at which it scores 2 or higher on the scale below. An identical feeder in afternoon sun gives the paired number, and a sealed jar from the same batch gives a third, since only temperature acts on it.

Why the existing evidence is thin#

The published intervals are sound and deliberately conservative, but they are rules of thumb with no open dataset behind them. The anecdote that exists rarely carries a thermometer at the feeder or a scoring definition written down in advance. "It went bad fast in the heat" is not a data point. "Score 2 at 41 hours, shaded feeder, daily max 88°F (31°C) at feeder height" is.

Know the limit before you start. Yeasts and bacteria multiply well before sugar water looks any different, which is why the cleaning guidance says clear nectar is not clean nectar. You are timing when an attentive person could have seen the problem, not when the nectar became unsuitable. No result here licenses leaving nectar out longer than the published table.

What you need#

• Two identical nectar feeders, same model and material, fully dismantlable. Saucer designs score most easily.
• One clear glass jar with a sealing lid, as the no-visit control.
• Plain white granulated sugar and ordinary tap water. Nothing else.
• A digital min/max thermometer, ideally two, with card to shade the sensor.
• A white index card to back the reservoir when judging clarity, and a phone camera used from a mark 12 in (30 cm) away.
• The kit from how to clean a feeder: bottle brush, port brush or pipe cleaners, unscented household bleach, a tub, gloves.
• A printed copy of the 0 to 4 scale.

How to set it up#

Pick two positions in the same yard at the same height, about 5 ft (1.5 m), at least 12 ft (3.7 m) apart so one territorial bird is less likely to guard both. One sits in full shade all day; the other takes 4 hours or more of direct afternoon sun.

Clean both feeders and the jar by the full method before every run: dismantle, scrub in hot water with the port brush, soak about 10 minutes in 1 part household bleach to 9 parts water, rinse thoroughly, dry completely. Mark a fill line on both at the same volume, 4 fl oz (120 ml) is plenty, since fill depth changes surface area and thermal mass.

Hang the min/max thermometer in shade at feeder height, sensor out of direct sun. One left in the sun reports only how hot the thermometer got.

The procedure#

1. Clean both feeders and the jar the evening before and dry them. In the morning mix one batch at the site standard: 1 cup of sugar (200 g) to 4 cups of water (about 950 ml), stirred until dissolved and cooled. Record whether you boiled the water, then do the same every run.
2. Fill both feeders to the marked line from that single batch, and the jar from it too, so the recipe cannot explain a difference between feeders.
3. Hang the feeders, seal the jar and set it in shade beside the shaded feeder. Photograph all three, log the fill time to the minute, reset the thermometer.
4. Score all three twice a day at fixed times, say 8 a.m. and 6 p.m., with the same order and observer all season.
5. At each check read the min and max, then reset. Hold the reservoir against the white card in open shade for clarity, look for strands and specks, tilt it for surface film, photograph it from the marked spot.
6. Smell last, wafting air toward you from about 6 in (15 cm) away. Do not put your nose over an open feeder and inhale.
7. End a feeder's run at the first check scoring 2 or higher, when the nectar falls below a scoreable depth, or at 7 days, whichever comes first.
8. When a run ends, take that feeder down at once: empty, scrub, soak, rinse, dry, then back in service with fresh nectar. Never leave it up to see how long it takes to reach 3. Only the sealed jar goes past 2, and it is where you learn what 3 and 4 look like.

What you record#

One row per unit per check, scored against the printed table, not from memory.

Keep the three failure modes separate. Cloudy is the first visible sign that something is multiplying, and it is your endpoint. Fermenting means yeast has taken hold and is turning sugar into alcohol and gas, hence the strands, bubbles and brewery smell. Molding means fungal colonies anchored in micro scratches, usually around the ports, where brushing will not clear them. Nectar can go cloudy without fermenting visibly, and ports can mold while the liquid looks fine, so score the components as well as the total.

Also record position, feeder and batch identity, whether the batch was boiled, fill volume, elapsed hours, min and max since the last reset, direct sun hours, rain, nectar remaining as a percentage of fill, bird use, the photo name, and why the run ended.

Controls, and the ways this goes wrong#

The sealed jar controls for contamination: same air temperature as the shaded feeder, no bird tongues, insects or airborne debris. The jar-to-shade difference is roughly the contribution of use, the shade-to-sun difference that of solar heating.

Swap the two feeders between positions on alternate runs, or feeder identity rides along with position and one scratched reservoir looks like a sun effect. Retire any feeder gone cloudy or crazed, or with established port mold, because mold anchored in scratches seeds the next run.

Some things move the score without being spoilage: rain through the ports, spring pollen, dead insects, a dusty week. Flag those runs rather than scoring around them. Observer drift is the quiet killer, since the haze you called 1 in June becomes a 2 in August once you expect it. Use the printed table, keep one observer, photograph every check.

How many runs before the number means anything#

One run is a story about one weekend. Target 12 paired runs across the season, and report by the site's existing bands: above 90°F (32°C), 85 to 90°F (29 to 32°C), 75 to 84°F (24 to 29°C), 65 to 74°F (18 to 23°C), 55 to 64°F (13 to 18°C), and below 55°F (13°C). Assign a run by the mean of its daily maxima at the shaded thermometer, and summarize a band only once it holds 6 runs; below that, publish rows and no summary.

Use the median hours to score 2, not the mean, and give the range. Runs that ended because the feeder drained, rain got in, or you hit the 7 day stop are censored rather than failed: report them as "at least N hours" and count them separately. Twelve runs in which nothing reached the endpoint before the published interval is a real result, and should be reported that plainly.

Submitting your results#

Use the datasheet for this protocol, one row per unit per check, including checks that scored 0. The zeros carry as much information as the endpoint.

Send completed runs, including partial and censored ones, through the Feeder Log. Include your rough location, the feeder model, where the thermometer hung, and your photos if you can share them.