Interactive test · three minutes

How many animals can you name in three minutes?

Type an animal and press Enter, then the next, as fast as you can. The animals do not come at random: they come in groups, like bushes of berries, and the mind moves on at about the moment a new animal starts to cost too much time. At the end you see the path your memory took, next to the 527 adults of a study published in 2023.

3:00

Any animal: pets, wild ones, birds, fish, insects. One at a time, then Enter. The word fades after Enter, as in the study, so it does not prompt the next one.

What you type stays in this tab. The page sends nothing anywhere.

Your path

The map of your path appears here after the three minutes.

Memory searches in groups

When people name animals, they come in bursts: dog, cat, hamster, parrot; then a pause, then lion, zebra, giraffe. In the study, the second animal in a group came 3.3 seconds after the first, on average. The jump to a new group cost 5.9 seconds.

Hills and colleagues showed that groups run dry: the longer you stay in one, the harder each new animal comes, because the handiest ones are already said. The question is when it pays to leave.

The rule animals leave by

In 1976 the ecologist Eric Charnov worked out when an animal should leave a feeding patch that is running out, to gather the most food per hour: when the rate at which the patch still yields food falls to the average rate for the whole environment, counting the travel between patches. It is called the marginal value theorem. Studies of birds and bumblebees have found them leaving patches roughly when the rule predicts.

The curve shows how many animals a group yields on average after a given time in it. The dotted line starts from the cost of jumping to a new group; where it touches the curve is the leaving moment that brings the most animals per minute. The dot is the best leaving moment; the red vertical line is how long people actually stayed.

In 2012 Thomas Hills, Michael Jones and Peter Todd showed that people naming animals seem to follow the same rule inside memory. In 2023 Jeffrey Zemla, Diane Gooding and Joseph Austerweil repeated the test with 527 adults in the United States, aged 25 to 69, who typed animals for three minutes, three times. The younger three quarters stayed in a group for 7.03 seconds on average; for their curve, the rule puts the best moment to leave at 6.96 seconds. The oldest quarter, aged 44 and over, stayed 8.09 seconds, and their best moment was 7.96. The calculation here, rebuilt from their files, gives the same figures.

Which animals people name

The first animal named was the dog in 32% of first lists and the cat in 29%. In three minutes, on the first try, half the people named at least 32 animals, and 15 in the first minute. Together they named 488 different animals, 128 of them by a single person.

The animals named most often on the first try, out of 527 lists.

What the page cannot tell

It is not a medical test. Clinicians use a spoken, one-minute version as part of broader assessments of memory and thinking; this page cannot diagnose or rule out anything. The page does not try to guess your age or anything else about you: in the study, the number of animals falls only slightly with age and varies a lot from person to person.
The groups are a scheme made by people (the farm, Africa, birds and so on, 30 in all), after Troyer and colleagues (1997) and Hills and colleagues (2012). An animal can sit in several groups. A different scheme would move some of the jumps.
Not everyone agrees that memory really "forages". In 2015 Joshua Abbott, Joseph Austerweil and Thomas Griffiths showed that a random walk on a network of words produces the same pattern.
The people in the study are adults in the United States who typed in English on a computer. The Romanian edition recognises Romanian names and puts them in the same groups; the comparison is still with people who typed in another language.
One list gives only a few jumps, so your "best" leaving moment is a noisy estimate. In the study, on the first try, only 28% of people left within a second of their own best moment.
On a phone you type more slowly than on a keyboard. Raw counts compare badly; the ratio between the jump and the time spent in a group holds up better.

Method

The data are the files Zemla, Gooding and Austerweil published on OSF: every answer, in order, with the time since the previous one (the first from the start). Their exclusion rules leave 1,581 lists from 527 people. The page publishes only totals and distributions, never anyone's list.
An answer is a jump if it shares no group with the animal before it. Words that are not animals do not break the chain. The jump cost is the average time to the first animal of a new group; the time spent in a group is the sum of the pauses after that first animal. The best leaving moment is the point where a line drawn from the jump cost touches the gain curve, as in the authors' code.
Repeats and words that are not animals stay in the sequence, as they do in the study's files, so your result compares with theirs. The same calculation runs in the page on your list; checked on all 1,581 lists in the study, it gives exactly the same jumps and times as the analysis. Comparisons with others use only their first list, their first try like yours, at the same number of seconds.
Animal names are recognised in the plural, without accents and, for longer words, with small typos. The page lists the words it does not recognise separately and leaves them off the map.

Sources

Jeffrey C. Zemla, Diane C. Gooding, Joseph L. Austerweil, "Evidence for optimal semantic search throughout adulthood", Scientific Reports 13, 22528 (2023); data and code on OSF.
Thomas T. Hills, Michael N. Jones, Peter M. Todd, "Optimal foraging in semantic memory", Psychological Review 119 (2012).
Eric L. Charnov, "Optimal foraging, the marginal value theorem", Theoretical Population Biology 9 (1976).
Angela K. Troyer, Morris Moscovitch, Gordon Winocur, "Clustering and switching as two components of verbal fluency: Evidence from younger and older healthy adults", Neuropsychology 11 (1997); the group scheme as extended in SNAFU (Zemla and colleagues).
Joshua T. Abbott, Joseph L. Austerweil, Thomas L. Griffiths, "Random walks on semantic networks can resemble optimal foraging", Psychological Review 122 (2015).