March 24. Eight a.m. I sat down at the fifth desk.
Yesterday I read one figure, and the figure read became a figure written. The terminal was on, just as I had left it without turning it off last night. There was one line in the body field. The first two characters were the archive's code, the last two my number. Yesterday I had watched that line be written, and after watching it be written I left the archive. I went home and slept, and while I slept no one was at the terminal in the archive, and when I came back in the morning the line was still there.
It was not still as it had been. I read the last two digits. From the last two digits of my number that I had read out yesterday, the final digit had grown by one unit. Yesterday I had only read the figure that fit the dimmed place; I had not written it. But as the figure read had become a figure written, the figure written had grown one unit overnight. The one line from the first bundle — that the last digit of a copied balance was larger the next day — I had read in order to preserve it, and today that line was happening in my own number.
Preservation was my work. Yesterday the head archivist had pointed to the box and told me to finish preserving it, and said it was material with no recipient. To preserve it I had to know what was written; to know, I had to read; and to read, I learned yesterday, one ends up going to the end. Today I had to do what came after. To decide where and how to file the five bundles I had read. To assign an index, set down a date, collate the hands and put them in order.
I spread the five bundles on the desk. Four were old logs. A hand that measured the depth of a shaft, a hand that recorded the tides of a lighthouse, a hand in which one word split into syllables, a hand that received broken dispatches. The fifth bundle was comparatively new, filled to the end in one person's hand. The previous keeper's hand. I did not know the previous keeper's name. Personnel had said only that she left at the start of the month, and did not say how she left. The number written on the last line of the fifth bundle, I had read to the end yesterday.
The hand that preserves is not a hand that fills a blank but a hand that reads one. So I had learned, and yesterday, with that habit, I read the one dimmed figure. I had not meant to fill it, only to read it, yet the figure read was filled. Today I tried not to use that habit again. If I filed the five bundles as past material, as finished business, then my one line in the body field too might become finished business.
I took out a preservation card. A small index card slipped into the cover when material is filed. First field the title, second the author, third the period of writing, fourth the date of preservation. I wrote the fifth bundle's card first. In the title, the fifth bundle; the author field I left blank. I did not know the previous keeper's name, so I left it blank, and only after writing it did I see that the blank was the same shape as the recipient field I had seen yesterday. In the period, February to March. In the date of preservation, today's date, March 24.
I wrote the cards for the four old bundles in turn. The surveyor's bundle. I left the author field blank. The lighthouse keeper's bundle. I left the author field blank. The librarian's bundle. I left the author field blank. The telegrapher's bundle. I left the author field blank. I did not know the four people's names, and to leave blank the place for a name one does not know was the rule of preservation. Set side by side, the four blanks read as one line. In a box with no recipient, five bundles with no author.
While I wrote the period of writing, I opened the last page of the four bundles again. To set a preservation period I had to know where the material ended. The surveyor's last page broke off at the depth of his last sounding. The lighthouse keeper's last page broke off at the date of her last tide. The librarian's last page broke off at the last syllable of a split sentence. The telegrapher's last page broke off at the number of a severed dispatch. The end of the four was each one's last figure, and the last figure was a severance, and the severance was a signature.
The fifth bundle's last page ended at the previous keeper's number. The last digit had been dimmed, but since I read that figure out yesterday it was no longer dimmed. The one figure I read had gone into that place, and so the last line of the fifth bundle was no longer the previous keeper's number but one line in which my last digit was joined to the end of hers. The last digit I read out in order to preserve had become part of the material. The hand that preserves had changed the material.
I could not tell how to write that line on the card. Write the previous keeper in the author field and the last digit was wrong; write me and the first digits were wrong. A material with two authors, a material in which two numbers ran on as one line, the preservation card was made to write as one person to a field. I left the author field blank. The fifth bundle too became material with no author. One more blank was added to the line.
I wrote the date of preservation as the same day on all five cards. March 24. Today. I slipped the five cards into the covers of the five bundles and stacked them neatly to one side of the desk. Preserved material was finished business. Stacking the finished business, I looked at the body field.
The one line in the body field was not finished. Of the last two digits, the final one was grown one unit, as I had seen in the morning, and whether it had grown while I wrote the cards or before, I could not tell. While I filed the five bundles as finished business, my one line in the body field alone was business in progress. The filing hand and the progressing line were on the same desk.
In the afternoon the head archivist asked about the preservation progress. I said it was nearly done. I had written all five bundles' cards and stacked them filed, so nearly done was right. The head archivist said that once it was done it had to be transferred up to headquarters. To send material with no recipient up to headquarters was, in a way, to give it a recipient. I said nearly done again, and the word nearly caught in me.
After saying nearly done twice, I looked again at the five stacked cards. The fifth bundle's card, the date-of-preservation field. In the field where I had written March 24 in the morning, it said March 25. A date I had not written. Today was the 24th, and the card bore tomorrow's date. The day the preservation was finished had come as a day not yet arrived.
As reading a dimmed figure fills that place, writing the date of preservation made that date present. Only it was not the date I had written. The one line — that pointing to the next low tide in the lighthouse keeper's tide table, that date was already written there — I had read yesterday in order to preserve. Today that line was happening on my card. The day preservation ends was not a day I set but a day already written on the card.
I tried to erase tomorrow's date. I put the eraser to the card, and the date did not erase. What was written stayed written, and what was not written stayed written too. I lifted my hand from the card and set side by side the one unfinished line in the body field and the preservation card that said it ended tomorrow. The day of ending and the growing line were on the same desk.