IN THE INTERPRETATION of epigraphical texts, the partially preserved or indistinct letter may hold the key to the meaning of entire passages. Such letters are usually indicated nowadays by the use of a dot beneath. Since most of the stones on which inscriptions were cut are now broken or damaged, many letters are of necessity uncertain; and accordingly the use of such dots has become very general. An examination of certain problematical letters has tempted me to make a brief analysis of the whole subject of dotted letters, and to conclude that although in general there is agreement in definition, in practice there is disparity, and in some usages at least there is much danger. In the present study, I shall first briefly sketch the early history of the use of dotted letters. Second, I shall analyze the three functions of dotted letters in present-day Greek epigraphy. The third section, on which I lay the most emphasis, takes up in some detail, and with photographic illustrations, certain striking cases exemplifying two misusages of the dotted letter. Lastly, I propose measures which would help minimize the dangers arising from these misusages. In the publication of epigraphical texts in the nineteenth century it was customary to provide a majuscule text. The editor could indicate by the use of broken type just what he saw on the stone. For instance, if the vertical stroke of an epsilon was the only trace remaining, the editor printed an upright in the left side of the letter space. If the letter was very faint, but its identity unmistakable, the shape was indicated by a series of dots. Thus, a very indistinct omicron would be printed in the form O. This system was used for all Attic inscriptions published in the Inscriptiones Graecae. The practice of using dotted letters was apparently taken over from papyrology. This use of dots in papyrological publication was well enough establi hed by 1898 so that dotted letters were included in the table of sigla in the first volume of the Oxyrhy chus papyri. Grenfell and Hunt stated, Letters with dots under them are to be considered uncertain.'