More than 100 primary root segments from dormant seeds, 100 radicular segments from immature embryos, 450 whole immature embryos, and a number of whole seedlings of C. gronovii were studied under in vitro cultivation in the presence of such growth-regulating factors as: coconut milk, casein hydrolysate, auxins, kinetin, adenine, gibberellic acid, and additional B vitamins as supplements to White's medium. In normal embryogeny early swelling of the basal pole of the embryo and profuse production of epidermal hairs upon germination are superficial characteristics of Cuscuta which resemble seedling root growth in other plants. At no time is there evidence of a primary-root meristem, nor at any time do cotyledons develop. It was not possible to induce root formation experimentally from any of the starting embryonic materials, nor from stems which were cultured to maturity culminating in the production of flowers after 6 months of in vitro cultivation, nor from callus derived from the radicular pole of the embryo. The undifferentiated radicular pole frequently gave rise to rapidly growing callus from which new shoots were formed spontaneously, although some immature embryos completed normal embryogeny. Embryo size and subsequent type of development were not strictly correlated but there was a tendency toward the following patterns. Embryos less than 0.4 mm in length lost their original form, callused evenly throughout and became masked by development of buds over the entire surface. Embryos 0.4–1.2 mm were most apt to retain their original form; callus developed at the radicular pole but not at the shoot pole. Embryos 1.5 mm and larger frequently gave rise to radicular callus but also grew normal shoots. Embryo inoculants which had begun to coil were most apt to complete normal embryogenesis. No correlation could be drawn between the addition of specific growth regulators or combinations of regulators and specific patterns of development. It is suggested that in the evolution of Cuscuta from autotrophic ancestors, the loss of both cotyledons and roots was a single embryonic event and so complete that even under the influence of a highly modified biochemical environment it would be impossible to derive roots from dodder tissue. The usual interpretation of the dodder haustorium as a highly modified adventitious root in all probability is not valid. Is is further suggested, however, that this loss of root and cotyledon is not a necessary consequence of the evolution of the parasitic mode of nutrition, nor of the gross adult growth form.