When considering the huge work of the <i>Naturalis Historia,</i> one wonders how Pliny was able to handle the more than 34,000 items he had to work with. The excavation of the writing tablets of Vindolanda sheds new light on this question, after O. Schaaber had suggested the use of descriptors (keywords). The present authors assume the following procedure: a servant read out of a book and Pliny dictated parts of the text to another servant who wrote it down onto wax tablets and Pliny added short keywords. The next step was writing down these fractures of the original text onto the thin wood tablets, using one for each item, and adding the keywords. Then it was easy to order the tablets according to the subjects. From these assorted tablets the <i>commentarii</i> were made, and, finally, the books could be written. -- AATA