September 13. What comes in has passed what goes out.
At twenty past nine I opened the post.
Three envelopes. All say they come from the side that came in by the outside reading.
In the first envelope there are six copied pages.
A tide table. The same form as sub-number three.
The year is thirty-five.
At twenty to ten I opened the second.
Two typed sheets.
A list of station numbers, it says. The same head as in sub-number two.
Twenty-nine numbers. Four of them are not in the original.
The soundings of the four exceed the largest value in the original.
At ten I opened the third.
One sheet. The form of a register of messages.
At the top right there is a page number.
Page one hundred and fifty-five.
Sub-number four ends at a hundred and fifty-four.
At half past ten I counted what had gone out.
One reply in March. One reply on the second of August.
Two.
I counted what had come in. Three inquiries and three envelopes.
Six.
At twenty to eleven I looked at why the count only piles up on one side.
What goes out has no reply field. Once out, nothing remains.
What comes in comes on paper. Paper remains.
The only basis for counting what went out is my log. The basis for counting what came in is on the desk.
From eleven to twelve I looked for where to put what had come.
There is a donation procedure. Five boxes.
Date of donation, name of article, quantity, donor, confirming officer.
The donor box takes a name, it says.
All three are eight Latin characters. That is all the envelopes carry too.
At ten past twelve I looked at the procedure.
Where a name cannot be ascertained, enter unknown, it says.
In February I amended four to unknown under the same procedure.
I entered three as unknown.
At half past twelve I opened the list of persons entering.
The date last entered for the unknown row was thirty-seven.
It is the thirteenth of September, two thousand.
At one I applied the sort.
The compiler row is first. Its date is blank.
The unknown row is last. Its date is today.
A row that stood high for sixty-three years went to the bottom today.
The row moved because what is entered in it grew.
From half past one to two I matched what had come against the originals.
The six tide table pages are in the same form. The rule lines are the same weight and the tide-name system is the same.
Sub-number three ends in thirty-four. Why it ends is not in the original.
The four station numbers are not in the original. The numbers before and after them are.
The one register page is the hundred and fifty-fifth. It follows the hundred and fifty-fourth.
All three attach behind an original.
From two to twenty to three I traced back what has grown.
A row was gained, a value grew, an empty box was filled, a date moved, a page was gained.
All of it grew in a copy. It was inside this room.
What grew today came from outside.
At twenty to three I looked at the envelopes again.
There is no address. The sender box is empty.
There are postmarks.
All three are different post offices.
The account rendering is the same in all three. Eight Latin characters.
At twenty past three I opened the query record.
I looked at what that account has opened until now.
Title, imprint, the eighth field.
Nothing came in on the word thirty-five. That year is not in this room's materials.
Nor is it in anything that left this room.
From four to twenty past four I wrote out where it could have come from.
What left this room is the title, the imprint, the eighth field and two replies.
Thirty-five is not in any of them. Nor are station numbers. Nor a hundred and fifty-fifth page.
It was not told from here.
What is not here was sent here.
At half past four I opened the right-hand material.
Four hundred and seventy-four pages. The date on the page gained is the twelfth of October of ninety-eight.
My log is also on its four hundred and seventy-fourth page today.
At ten to five I opened the box.
Seven. Six impressions. The two hundred and second day.
I did not put the three in. They are on the desk, not yet registered.
At five I looked at the sixth box.
Two thousand six hundred and twenty-four.
At ten past five I ran off the query record. The set's row was opened nineteen times yesterday. Nine by the imprint route.
At a quarter past five I looked at the confirming officer box.
It is the fifth box of the donation form. It takes the name of the confirming officer.
There is one person in this room.
The donor field reads unknown and the confirming officer field reads my name.
Not known and known stand side by side in one row. In February too there were two unknowns in one row.
This time one is unknown and one is a name.
At half past five I looked at the hundred and fifty-fifth page again.
One line is entered. Three boxes.
The time received box is empty.
The time sent box carries the thirteenth of September, two thousand.
Today.
At twenty to six I read the substance box.
This bundle passes to whoever next reads this line.
It is the line I wrote in the note field on the fifth of July.
It showed in the eighth field on the screen, so it did go out.
What went out has come back, entered on the hundred and fifty-fifth page of a form from thirty-seven.
I have written all along that there is no reply field.
This did not come as a reply.