Years ago (16, 18, 19), in a study of shadowed preparations of Proteus vulgaris that had been autolyzed in the cold, the observation was made that the flagella arose from basal bodies. However, recently (3, 7, 24, 33) doubt has been cast on the conclusion that the flagella of bacteria emerge from sizable basal bodies. This problem has, therefore, been reinvestigated with actively developing cultures of Proteus mirabilis, the cell walls of which had been expanded slightly by exposure to penicillin. Two techniques were applied: ultramicrotomy, and negative staining of whole mount preparations. This paper deals with the thin sections of bacteria after the usual fixation technique had been altered slightly: the cells were embedded in agar prior to their fixation and further processing. The flagella then remained attached to the cells and were seen to extend between the cell wall and the plasma membrane. Occasionally, the flagella appeared to be anchored in the cell by means of a hook-shaped ending. In sections of cells rich in cytoplasm, the basal bodies are particularly difficult to visualize due to their small size (25 to 45 mmicro) and the lack of properties that would enable one to distinguish them from the ribonucleoprotein structures; in addition, their boundary appears to be delicate. However, when the cytoplasm is sparse in the cells, either naturally or as a result of osmotic shocking in distilled water, the flagella can be observed to emerge from rounded structures approximately 25 to 45 mmicro wide. Contrary to a previous suggestion (21), the flagella do not terminate in the peripheral sites of reduced tellurite, i.e. the chondrioids. The observations in this part of the study agree with those described in the following paper (15) dealing with negatively stained preparations.