The Ethogram

Pick one animal you can observe safely and repeatedly. Watch it for a full twenty minutes without recording anything.

This is not wasted time. You are finding out what the animal actually does, rather than what you expected.

An ethogram written before you have watched will contain categories the animal never performs and miss the ones it performs constantly.

List 6-12 behaviours. Give each a short name and a one-sentence definition using only observable body position and movement.

Test every definition against this question: could a stranger apply it without knowing what I think the animal feels?

Reject "curious", "nervous", "bored". Keep "head below shoulder line, beak in contact with substrate".

Every interpretive word you leave in is a place where two observers will disagree.

Check every pair of behaviours: can the animal be doing both at once? If so, split them or nest one inside the other.

Then add an explicit "Out of sight" and an explicit "Other" category.

At any instant, exactly one category must apply.

Without "Out of sight" people silently record the last thing they saw, which invents data — and it is the single commonest fault in amateur behaviour records.

Set a timer to beep every 30 seconds for 30 minutes. At each beep, record the single category the animal is in at that instant — not what it did in between.

You now have 60 samples.

Count each category and divide by 60 to get its share of time. That is the behaviour budget.

Recording at the beep, and only at the beep, is the whole method. The moment you start recording the interesting things between beeps, the budget is biased towards rare, dramatic behaviour.

Have a second person score the same animal at the same beeps, independently, using only your written ethogram.

Count the samples where you agreed, and divide by the total.

Expect well below 100 % on the first attempt. Find the categories you disagreed on, rewrite those definitions, and repeat.

This is inter-observer reliability, and it is the measurement that decides whether your ethogram is a scientific instrument or a private diary. Rewriting the definition is the fix — arguing about who was right is not.

Run the same 30-minute scan sample in two conditions — morning against afternoon, warm against cold, before food against after.

Plot the two budgets side by side as stacked bars.

Expect some categories to shift substantially and others not at all.

An ethogram measured once is a description; measured twice under different conditions it becomes a comparison, and a comparison is the smallest thing that can test an idea.

Ethology's founding argument was that behaviour is a biological trait like any other — it can be catalogued, compared between species, and asked evolutionary questions about, exactly as a bone or a beak can. That was not obvious. The dominant alternative in the early twentieth century was laboratory behaviourism, which studied a few species in artificial settings and was largely uninterested in what an animal does in the wild.

Tinbergen's four questions, set out in On Aims and Methods of Ethology (1963), are the framework that made the field coherent. For any behaviour you can ask four separate things: what causes it right now (mechanism); how it develops over the animal's lifetime (ontogeny); what it is for (adaptive function); and how it evolved (phylogeny). His central point was that these are not rival explanations — they are four different questions, all legitimate, and arguments in animal behaviour are very often two people answering different ones and thinking they disagree. Tinbergen, Konrad Lorenz and Karl von Frisch shared the 1973 Nobel Prize in Physiology or Medicine for founding the discipline.

Sampling rules were the other half of making it rigorous. Deciding in advance when you will look — scan sampling at fixed intervals, focal sampling of one named individual, or all-occurrences recording for a rare event — is what removes the observer's attention from the results. Ad libitum watching still has a use, but only for building the ethogram in step 1, never for producing the numbers.

Honest limits. An ethogram is species-specific and often population-specific; one written for garden blue tits will not transfer to a parrot. It records what an animal does, never why, and the temptation to slide from a clean behavioural category back into a motive is constant — this is anthropomorphism, and defining categories physically is the guard against it. Rare but important behaviours are badly served by interval sampling and need all-occurrences recording alongside. Observer presence changes behaviour, so a habituation period matters. And the most important limit is ethical: observe, do not interfere. Do not bait, handle, corner or disturb a wild animal, do not approach nests or dens, and stop if the animal changes its behaviour because of you — at that point you are no longer measuring the thing you came to measure.