$21,400 Left A Treasury Account In May 1955. Nobody Recorded What It Bought.

COLLEGE PARK, Md. — Twenty-one thousand four hundred dollars. Fiscal year 1955. One subproject number, one conduit foundation, one authorising initial, one date.

That is the whole of the line. It is one row of a schedule of two hundred and six rows, and it is the single most quoted piece of financial evidence about the programme it belongs to.

We took it to an archivist who has spent twenty years in that record group, and to a forensic accountant who rebuilds federal disbursements of this vintage for probate courts, and asked a narrow question: what is a document of this type physically able to establish?

Four things it will bear

• That a sum was authorised. $21,400, entered in figures and in words, the two agreeing.
• That it was authorised on a date. 12 May 1955, in the schedule's own sequence, with the rows either side of it dated consistently.
• That a named signature authorised it. Two initials in the approving block, above a typed title.
• That it was payable to a named intermediary. A foundation, with an address, appearing as payee on the face of the form.

Those four are solid. They are also, as Ivor Brandtsen-Muriuki puts it, four facts about a form rather than four facts about the world past the payee line.

Five things it will not

• What was bought.
• Whether the money was ever spent.
• Who finally received it.
• Whether whoever received it knew where it had come from.
• Whether the subproject number on the voucher describes the activity the money funded.

The fifth one undoes most published readings

Subproject numbers were cost centres. They were opened for accounting convenience, not to describe work.

Delphine Ostrowski-Bayard keeps a working index of the anomalies. She showed us forty-one numbers that appear in the finance record with no corresponding administrative trace anywhere; nine cases where one number carries two plainly unrelated payees inside a single quarter; four numbers reopened years after closure, and reused; and two split mid-year so that one number became two, both of which continued to be quoted afterwards as though each were a distinct enterprise.

A number is a pocket. It tells you which pocket the money came out of. It does not tell you what the money did.

Three times we did this ourselves

We went back through our own published work and counted. These are the three worst.

In They Burned The Files And Kept The Receipts we reported that eleven cost centres appear only in the voucher record. That is true and we stand behind it. But we never wrote the sentence saying a cost centre is not an activity, and readers took eleven numbers to mean eleven hidden programmes, which is not established and which we did not correct for four months.

In The Fluoride Contract Went Through On The Second Try we set an award figure and a policy change in consecutive sentences and let the running order imply a purchase that no source in the story alleged.

In Six Firms Sold Him The Harvest we described a subsidy line as though someone had decided to subsidise that particular sale. The subsidy was a published schedule. It applied to anybody who shipped.

Why the file is silent

There is no smoking gun in the financial record, and the reason is not that somebody scrubbed it.

The disbursement system running in 1955 predates project accounting entirely. It exists to satisfy an auditor of two things: that a sum left a Treasury account, and that it reached the payee named on the form. It does that job perfectly. It has no field for purpose because purpose was never the auditor's question — that belonged to the appropriation, and appropriations are drafted in language wide enough to cover a decade of whatever anyone thinks of next.

On this form, a payment to a laboratory and a payment for six months of a department's stationery are the same document.

What this desk takes from it

A voucher line is a receipt. It survives because receipts are kept in a different building under a different retention schedule from the papers people actually burn.

Professor Ines Karadžić-Bell's reply is printed below in full, and it is the most serious objection this piece received: that the absence of a purpose field is not a fact of nature but a consequence of a conduit structure somebody chose, and that a newspaper which prides itself on deflation can end up reaching for one as reflexively as another paper reaches for a revelation.

We have not answered her. We have printed her, and we have printed the sentence we could not get past: that none of what is known about these laboratories, from testimony and litigation and the records that were not destroyed, turns on $21,400 at all.