Although the original Glagolitic alphabet is nowhere preserved in its entirety, considerable evidence remains from which it can be reconstructed. Since this evidence is not self-consistent, it must be weighed, not merely counted, and some of it must be discarded. Previous scholarship has erred in its weighing of some of this evidence, in particular by overvaluing the evidence of early Glagolitic and Cyrillic abecedaria. The reconstruction offered here rests primarily on the attested numerical values of Glagolitic letters in early manuscripts and on the acrostic in an early poem by Constantine of Preslav, supplemented to a very small extent by Monk Xrabr’s treatise “On the Letters” and by the abecedaria. (Two new emendations are also proposed in the text of the acrostic poem.) The present reconstruction gives reason to challenge the common opinion that Constantine of Thessalonica, who created the Glagolitic alphabet, did so on the basis of a sophisticated and accurate phonological analysis of some Slavic dialect. It also gives reason to suppose that Constantine had access to information about the Armenian alphabet and its creator, Mesrop-Maštoc‘, and that he might have taken the Armenian alphabet as a model for the choice and the order of the letters in the latter part of his original Glagolitic alphabet.