The morphology of Proto-Indo-European has traditionally been reconstructed as a composite of all the complexities found in the daughter languages, with a full array of tenses, moods, and aspects. This dissertation re-examines that premise by focusing on one morphological category, the usdsusd-aorist, and its distribution in the Indo-European languages. An attempt was made to collect all forms of the usdsusd-aorist/preterite in five major languages (Rgvedic Sanskrit, Gathic Avestan, Old Church Slavonic, Old Latin, and Homeric Greek,) so that a more empirical analysis could be made. Other IE languages with usdsusd-constructions (Hittite, Tocharian, Armenian, and Old Irish) were also analyzed. It was found that, despite its apparent wide distribution, the usdsusd-aorist cannot be reconstructed as a productive category for the oldest layers of the proto-language. Rather, it must have developed primarily in "late Indo-European," in the eastern area (i.e. in Indo-Iranian and Greek), like a number of other morphological categories (the imperfect tense, the long-vowel subjunctive, etc.). On the other hand, the languages of the western area which had usdsusd-preterites (Italic and Celtic), as well as Hittite and Tocharian, show unsystematic or very limited usdsusd-constructions, or else the forms usually considered to be reflexes of the usdsusd-aorist actually owe their existence to other constructions. These findings support the "Space-Time Hypothesis" first proposed by Meid (1975), which characterizes the proto-language as a dynamic, developing entity, one which shows a distinct geographical split between east and west in its later stages. The western languages and Hittite, to the extent that they contain reflexes of the usdsusd-construction at all, represent an archaic layer of development; the eastern languages, with their more fully developed usdsusd-construction, innovated in fitting this form into the aorist system. A further important conclusion which was reached concerns the connection of the lengthened grade to the usdsusd-aorist. Traditionally, the L.G. is assumed to be reconstructible for a PIE usdsusd-aorist. However, in each language where it is thought to exist, the L.G. is shown to have developed secondarily and independently through the operation of compensatory lengthening. The lack of a fixed vocalism for the usdsusd-aorist reinforces the main conclusion of the dissertation, that the usdsusd-aorist was not a productive category of the old Indo-European.