muihuong.vnScent keeps what you forget.
Menu
Tiếng Việt

Learn · Part One — Learning to smell · Lesson 9 of 9

First rain is the smell of bacteria

Rain has no smell. What you smell was already lying in the soil, waiting to be thrown up into the air.

Everyone recognises the smell of the first rain of the season. Nobody has to learn it. And almost nobody has a word of their own for it — it takes a whole clause to describe.

The remarkable thing is that the smell does not come from the rain.

Where it comes from

The main compound responsible is geosmin, secreted by bacteria of the genus Streptomyces living in the soil. It sits there throughout the dry season.

When a raindrop hits dry ground, it traps tiny air bubbles in the film of water, and those bubbles burst, flinging up tiny aerosol particles carrying geosmin into the air. The mechanism was filmed with high-speed cameras in 2015 and looks exactly like a tiny explosion.

So the rain does not create the smell. It is only the means of spreading it.

Why everyone recognises it

The human detection threshold for geosmin lies at a few parts per trillion — studies report figures from four to one hundred. It is one of the lowest thresholds ever measured for human smell.

You will come across a popular comparison here: that the human nose is two hundred thousand times more sensitive than a shark is to blood. The comparison does not stand up, because humans smell in air and sharks smell in water — the two media cannot be converted directly into each other. We go into this in more detail in the first-rain dictionary entry.

The widely accepted hypothesis is that this has survival value. Geosmin points to water, and for ancestors living in dry regions that information was worth a life. Natural selection pushed this sensitivity very high and kept it there.

One compound, three settings

Geosmin is also what makes tap water taste muddy during algal blooms, what makes fish from stagnant ponds taste of mud, and what gives beetroot its earthy smell.

Three phenomena that sound unrelated, one molecule.

This illustrates the principle from the lesson on sticky rice and pandan: a smell belongs to a molecule, not to its source. Once you recognise geosmin in first rain, you will recognise it in stone moss, in leaf litter, and in a glass of tap water some morning.

The story of a name

This smell has had a name of its own only since 1964, when two Australian scientists coined the word petrichor, built from Greek roots meaning roughly the blood of stone.

The lack of a word for so many centuries does not mean a lack of experience — it only means nobody had felt the need to name it.

That is also why this dictionary exists.

Further reading for this part

A shared list for the whole of Part One — Learning to smell, not a note for every sentence. We state what each source was used for.

  1. Asifa Majid and Niclas Burenhult, 2014. Odors are expressible in language, as long as you speak the right one A study of the Jahai of the Malay Peninsula, a group with an abstract smell vocabulary. It underpins the central argument of Part One: the gap lies in vocabulary, not in the nose.
  2. Caroline Bushdid et al., 2014. Humans can discriminate more than 1 trillion olfactory stimuli The trillion figure is widely quoted. We cite it together with its critique, because Meister (2015) and Gerkin and Castro (2015) showed problems with how it was extrapolated. It is our example of how an attractive number can spread without anyone checking it.
  3. Những cây thuốc và vị thuốc Việt Nam A Vietnamese-language source for the species names, parts used and sensory descriptions of most of the medicinal plants in the dictionary. Where it and other literature disagree — as with garden mint — we give both.

One thing must be said plainly. We have not read every work above in the original; most of the knowledge comes through secondary literature and reviews. We say so rather than let the list suggest a depth of research we have not reached.

Last updated: September 2026