A Body of Work

The FSA photographs are still remembered not because any one of them is exceptional, but because together they describe something no single frame could: what happened to rural America — and to the southern plains the dust storms crossed — in the 1930s.

So the question in this chapter is no longer whether a photograph is good. It is:

What is this one’s relationship to the ones I already have?

One is a sample, a set is a distribution

It is worth being precise about what a set adds, because it is not quantity.

A single photograph supplies a sample. Looking at the frame above, you know that this happened, in that place, on that day. What you cannot tell from it is whether this was a rare extreme or a sustained condition. The photograph does not carry that information.

A set supplies a distribution. Thirty frames, different farms, different months, different families, all resembling each other — now “this is ongoing” is something the pictures say themselves, rather than something a caption asserts on their behalf.

This is a statistical intuition applied to looking: n = 1 cannot distinguish an accident from a norm. The persuasive power of the FSA archive rests on that, not on any one frame being moving.

There is a second mechanism, more immediate than the first: repetition forces comparison. A viewer reaching the second photograph automatically measures it against the first; by the third they have built an internal sense of what is normal for this set. From then on difference becomes content — someone smiling, light arriving from another direction, things that carry no information in isolation now carry a great deal.

A useful way to feel the second mechanism: a photograph of a closed shop is a fact. Twenty photographs of closed shops on one street, made over a year, is an argument — and nothing in any individual frame had to change for that to happen. The argument was built out of the gaps between them.

The third mechanism is sequence. Photographs placed side by side get read as ordered, even when nothing connects them in time. That is a burden and a tool: whatever order you choose will be read as an argument.

A series is not an album

A hundred photographs put together is an album, not a series.

The definition is: these photographs share something, and that shared thing produces a meaning none of them has alone.

The shared thing can be:

• Place — one street, photographed for a year.
• Subject — only doors, only signage, only people waiting for buses.
• Time — the same hour, the same window, every day.
• Light — only in rain, only at blue hour.
• Form — all square, all monochrome, all one focal length.

The test is simple: remove the shared thing and what is left? If the answer is “several individually decent photographs”, it is still an album.

Note that the shared thing comes in two grades, and beginners usually stop at the first. The first is a category — “these are all doors”, “these were all made in rain”. It gives the set a boundary, but it is not content in itself, because it only answers what these are. The second is a claim — “the doors on this street are being replaced by roller shutters, one at a time”. That answers so what.

A category is the entry fee; a claim is the work. And the claim is rarely available at the start — it tends to surface around the thirtieth or fiftieth frame. Which is why a series cannot be thought out, only photographed out. You have to accumulate enough to see the pattern before the pattern can tell you what the set is about.

Constraint is a gift

Being told to make a series usually produces panic, because there is nothing obvious to photograph. The cure is to impose a constraint so narrow it feels absurd.

“I am going to photograph London” is a subject nobody can begin. “I am going to photograph every shadow on the walk from my flat to the station” can be started this afternoon.

The constraint does more than get you moving. Its real function is that when you cannot change the subject, the only remaining direction is deeper. In the first week you exhaust the obvious shadows. In the second you are forced into the edges of shadows, the colour inside them, the path one travels across a wall in a day. That being forced is the improvement.

Some photographs exist for the set

This is the hardest thing to accept, and close to the deepest difference between a series and a single image.

Not every frame in a set has to stand alone. Some are there to:

• Establish. Tell the viewer where we are and what the rules are. Usually flatter and wider, unremarkable alone, but without it the close frames happen nowhere.
• Connect. Two strong photographs adjacent will weaken each other; a quiet one between them strengthens both. Its value is entirely positional.
• Let the viewer breathe. A run of pictures needs somewhere to rest. Empty, dark or distant frames do that work.
• Provide a control. One ordinary frame exists so that the extraordinary one can be read as extraordinary.

The reverse holds too: some very strong single photographs belong to no set at all. They should live in a folder called “singles” and be published on their own rather than forced in.

The practical consequence is that judging a photograph requires first asking in what capacity. The same frame can be a C as a single and an A as the seventh in a sequence. Applying single-image standards to a series selection produces a pile of competing climaxes.

How many

There is no correct number, but there are working ranges, and they are set by the occasion of viewing rather than by the work:

Occasion	Roughly	Why
A social post or thread	6–10	Most people stop around the tenth
A small exhibition or web feature	15–25	About as long as a viewer will stand
A small book	30–60	Page turning supports a longer rhythm
“Where the long project is now”	Show 10	What people see is always a selection

“More is better” is wrong, and specifically so: adding a weak frame does not add zero, it subtracts — it lowers what the viewer expects from the set, and that expectation colours everything they see afterwards.

A useful self-test: if you could only show five, which five? Those five are usually the actual skeleton, and everything else is preparing for or recovering from them. If you cannot choose five, the set is not ready to be shown.

Editing matters more than photographing

A series of twenty is typically drawn from two hundred. Choosing decides more about the result than shooting did.

Three questions while choosing:

1. Does this one add anything? If it says what an already-selected frame says, keep the better one.
2. Does it work beside its neighbour? Sets have rhythm, and rhythm problems only become visible with two or three adjacent frames side by side. The next section is entirely about arranging it.
3. Would removing it weaken the set? If not, remove it.

The hardest cut is a photograph you love that does not belong here. Nothing is wrong with it; it belongs somewhere else. One practical device removes most of the pain: keep a folder called “good, but not this”. Put it there instead of deleting it. Most people are unwilling rather than unable, and giving the frame a destination makes the judgement accurate again.

Arranging the rhythm

Rhythm sounds vague and has directly operable rules:

Symptom	Usual cause	Fix
Interest dies in the middle	Several frames in a row at the same distance or composition	Alternate wide, middle, close; move the landing point side to side
Exhausting to get through	Every frame is dense, every frame a climax	Put quiet frames between strong ones; emptiness is content
Cannot tell what it is about	No opening	The first frame establishes place and rules; it need not be the strongest
Ends for no reason	No closing	The last frame must read as an ending — pulling back, going dark, someone leaving
Reads as two sets	One frame breaks the constraint	Find it; you usually already know which one

Do not open with your strongest frame. Everything after it will be a descent. The convention is to place it around two-thirds of the way through, with the earlier frames preparing for it and one or two afterwards to close.

One labour-saving method for sequencing: make small prints and lay them out on a table, or shrink them to thumbnails in a row on screen. At that size only the skeleton survives — masses of light and dark, density, direction of gaze — and every rhythm problem lives in exactly that layer. You cannot sequence by looking at one full-screen image at a time, because that scale shows you the quality of each frame rather than the relationship between them.

Six common symptoms and their causes

1. Each one is fine and together they have no force → no shared constraint; this is an album → find the commonest shared element among what you already have, declare it the constraint, and go back out to photograph to it.
2. There is a constraint and the set is repetitive → the constraint is too narrow, or every frame repeats one discovery at one distance → keep the constraint and vary distance and scale; a constraint governs the subject, not the method.
3. Three hundred frames and you cannot select twenty → you were photographing rather than building a set → select three candidates the same evening after each outing rather than facing three hundred at the end.
4. Selection stalls halfway → the criterion was never written down, so it drifts and the work becomes exhausting → write in one sentence what the set is about and keep it beside the screen; anything that does not serve it is out, however much you like it.
5. You show it and the response is “so?” → the set has a category but no claim → ask what the viewer should believe after seeing it; if you cannot say, it is not finished.
6. It starts feeling like filling a quota → the constraint is exhausted and the set is done → finishing a series is not a failure. Sequence it, give it an ending, and start again with a different constraint.

Punched negatives

The FSA photographic programme ran from 1935 to 1944 and left roughly 175,000 negatives, all now in the public domain — which is where this part’s photographs come from.

But the most instructive part of the programme is not what it photographed. It is how it decided what not to keep.

While editing, Roy Stryker took a hole punch to negatives he did not want. Those are known as the killed negatives, and many survive in the Library of Congress collection today with a punched hole through the middle of the frame.

Two things make this worth knowing.

First, it is editing in its most literal form. A punch is irreversible, so every hole is an explicit decision — and the sum of those decisions is the 1930s as we know it. What we see is not that decade; it is the decade Stryker selected.

Second, it is a reminder that editing is a form of power. Researchers examining the killed negatives have found frames that were not technically inferior, merely unhelpful to the story the programme was telling. This is not an accusation of fabrication — the photographs are all real — but selection shapes meaning, and selection always has a position.

Your own series is very much smaller and the mechanism is identical: the one hundred and eighty you cut decide what the set says, exactly as much as the twenty you kept.

An extreme case worth knowing: Bernd and Hilla Becher photographed industrial structures for close to fifty years, from 1959 until Bernd’s death in 2007 — water towers, blast furnaces, winding towers — always from the same height, square on, under the same even overcast light, then exhibited them in grids. The constraint is nearly mechanical, and it is precisely that narrowness which lets a viewer see, for the first time, how much these buildings differ from one another.

The contact sheet, and what it exposed

Before digital, the editing tool was the contact sheet: an entire roll printed at negative size on one page, thirty-six frames in a grid, examined with a loupe and marked in wax pencil.

It enforced two things that are easy to lose now. Every frame was seen at the same small size, which is exactly the scale at which rhythm and structure are legible. And every frame was seen in the order it was made, so the process was visible — the approach, the circling, the moment of getting it.

When Magnum began publishing photographers’ contact sheets, what they revealed was reassuring rather than intimidating: around the famous frame sit many that are almost right and clearly not it. The celebrated picture is nearly never the first exposure, and rarely the only good one. The difference between those photographers and a beginner is much less in the hit rate than in the ruthlessness of the marking-up afterwards.

The digital equivalent is not the full-screen review your camera offers. It is a grid of thumbnails, flagged fast, at a size too small to admire.

Starting your first series

A concrete opening:

1. Choose somewhere you pass every day — not somewhere you want to go and never have.
2. Impose a constraint (one focal length only, monochrome only, one colour only).
3. Photograph for thirty days, at least one frame a day.
4. On the thirty-first, lay out all thirty and choose ten.
5. Look at the ten. You will see your own preferences — what light you like, how far away you stand, what you look at.

Those preferences are where style begins. Style is not chosen; it is discovered after the fact.

A table for the field and the desk

Stage	The question	If you cannot answer
Before going out	What is the constraint today?	Set one, then go
Before raising the camera	How does this relate to what I have?	Do not press; keep walking
Same evening	Which three from today are most like this set?	Select now; never accumulate
Halfway through selecting	What is this set about, in one sentence?	Write it down before continuing
Sequencing	What does the first frame establish, what does the last one end?	Add an opening or a closing frame
After showing it	Did they describe it the way you meant it?	If not, re-sequence before blaming them