The Pursuit of Shadows

For thirty years Yozo Itoh photographed the sky above Japan. In 1967, he published the results. “Mr. Itoh is not a professional meteorologist,” explains his book’s foreword, “but an earnest observer of clouds.” A write-up in the Quarterly Journal of the Royal Meteorological Society affirmed this distinction, pointing out the scientific error in many captions. “We are uneasy about some unusual combinations of the Latin terminology,” notes the reviewer, indicating with some disappointment “cases of incorrect coding of the clouds.”

If clouds are coded, then British chemist Luke Howard originated the cipher. His 1803 “Essay on the Modifications of Clouds” begins: “Since the increased attention which has been given to Meteorology, the study of the various appearances of water suspended in the Atmosphere is become an interesting and even necessary branch of that pursuit.” In the essay, Howard proposes four general categories of cloud—cirrus, cumulus, stratus, nimbus—that remain foundational today.

This was a scientific breakthrough at the time. Howard’s nomenclature, and reputation, spread quickly, swept up in the surging current of taxonomy. Johann Wolfgang von Goethe, who was then minister to the Grand Duke of Saxe-Weimar, celebrated by penning “In Honor of Howard”:

That which no hand can reach, no hand can clasp,
He first has gain’d, first held with mental grasp.
Defin’d the doubtful, fix’d its limit-line,
And named it fitly. —Be the honour thine!

Nephology, the study and appreciation of clouds, blossomed, and by century’s end the International Cloud Atlas had cemented Howard’s contribution. Published by the World Meteorological Organization, the atlas is designed as an authoritative reference for pilots, ship captains, farmers, and whomever else is bound by livelihood, or curiosity, to the impulses of weather.


It is not immediately clear that such a thing—an atlas of clouds—should exist. Clouds come and go, shift and mimic. Hamlet, turning upward, describes to Polonius the camel-like cloud turned weasel-like turned whale-like. Moment to moment they can separate from haystacks into shoals of fish, then lengthen in ragged tendrils.

But it’s this inconstancy that spotlights Howard’s great insight, alongside the work of his contemporary Jean-Baptiste Lamarck, who devised a far less popular naming system in French. Though ephemeral objects, infinite in their nuance, clouds are not beyond classification. If their variation “were produced by the movements of the atmosphere alone, then indeed might the study of them be deemed an useless pursuit of shadows,” said Howard. But “the case is not so with Clouds.” Instead, they manifest in categorical shape and texture the invisible conditions around them.

In the decades after Howard’s essay, scientists scrutinized and expanded his idea. By 1891, the meteorology community had agreed on ten essential cloud forms, exponents of Howard’s original four—cirrostratus, for example, or stratocumulus. A flurry of publications, like William Clement Ley’s 1894 Cloudland, began the process of formalization. The International Meteorological Organization convened the surreally bureaucratic Cloud Committee, which posted a “request for cloud pictures” to illustrate the new taxonomy. They received hundreds of submissions from across the world, selected twenty-eight for publication, and paired the images with descriptive text in English, French, and German. Thus, in 1896, the first Cloud Atlas came to be.

Today’s atlas has 196 pictures. It is anchored in the same ten forms and includes a growing coda of specialty clouds, like the high nacreous bands that shimmer in arctic gloaming, their composition still unknown. The atlas is currently under revision by an expert task team that hopes to be finished in two years.

Neither clouds nor our naming of them has changed dramatically over time. If previous revision is any guide—years of deliberation between 1966 and 1975 led to the conclusion that fog and mist result from the same process—the substantive updates will consist of delicate edgework, no more. Core principles will not be overturned. Debate takes hold at the margins because at the margins is where the rule of taxonomy loses its firmness.

“It must be distinctly understood that many types of cloud merge imperceptibly into one another,” Cloudland says in passing. There’s the rub.

From Yozo Itoh’s Cloud Altas.

To good effect, we parse the continuity of experience with words, separate it into measurable bits. We use words to transcribe a language that is neither written nor spoken. The Linnaean formula, from kingdom down to species, helps us index a natural system that will never fully submit to indexation. While we note the scientific distinctions between clouds, those same clouds evolve and blend into each other. At what point do their names change? At what point do species differentiate? When does fog become mist? This is the limit of language—it is built upon approximation. The closer we peer at an essence, the more rapidly it eludes definition. At the margin, chaos beats us back.

Yozo Itoh’s anthology is also called Cloud Atlas, subtitled An Artist’s View of Living Cloud. It endeavors to be both a scientific and aesthetic contribution, filled with 206 photographs, four of them in color and the rest black and white. Captions provide technical cloud names along with metaphor: “clouds shaped like waterfall,” for instance.

Itoh, after all, is an earnest observer, not a meteorologist. He is an artist, not a scientist.

And while Itoh’s occasionally incorrect coding may generate unease among members of the Royal Meteorological Society, such narrow focus misses the point. Scientists and artists tackle the same eternal problem with different tools: they stamp order from disorder, they recognize that which endures amid constant change. We need the interpretive canvas of artwork as much as we do Latin taxa. These are complements, spinning incomplete stories, thread-by-thread patterning an infinite fabric.

Perhaps in a subtle nod to this fact, Itoh keeps two picture captions particularly terse. “Chaotic sky,” he writes, and nothing more. A soft admission of limits. The picture speaks for itself.

Dylan Walsh is a freelance writer based in New Haven, Connecticut. He writes about science, the environment, and social justice. Irregular updates on his work can be found at


  1. “It is better to have your head in the clouds, and know where you are… than to breathe the clearer atmosphere below them, and think that you are in paradise.”

    Henry David Thoreau

Commenting on this item is closed.