De Abusu linguae Graecae

As for the Plautine passage (which I also cited), nothing is to be believed less than that the poet, after having said symbolum, immediately repeated the same word in a different gender, saying symbolam. Rather, it happened by some miracle that the word which would not have escaped the eyes of any rash corrector in the preceding lines escaped them in those that followed, and at such a short interval. There remains the passage of Aulus Gellius (or Agellius), which is brought forward there as the firmest defense of the cause. For he closes the thirteenth chapter of the sixth book with these words: "Such were the symbola contributions at the house of Taurus, and others, which were of the second course, as he himself was accustomed to say, exagema delicacies." But I shall openly show that what can appear to be the firmest defense at first glance is very weak. For first, no one will deny that symbola could easily have been corrupted into symbola, just as in Plautus and Terence. Secondly, it is not obscure that such things can be retained, even if we read symbola. I shall certainly make it so that it is not obscure. For this is the beginning of that chapter: "This was practiced and observed at Athens by those who were closer to the philosopher Taurus. When he would invite us to his house, lest we should arrive entirely, as the saying goes, 'immune' and asymboloi without contribution..."