Although recent studies have indicated that the major histocompatibility complex–like, β2-microglobulin–associated CD1 molecules might function to present a novel chemical class of antigens, lipids and glycolipids, to α/β T cells, little is known about the T cell subsets that interact with CD1. A subset of CD1d-autoreactive, natural killer (NK)1.1 receptor–expressing α/β T cells has recently been identified. These cells, which include both CD4⁻CD8⁻ and CD4⁺ T cells, preferentially use an invariant Vα14-Jα281 T cell receptor (TCR) α chain paired with a Vβ8 TCR β chain in mice, or the homologous Vα24-JαQ/Vβ11 in humans. This cell subset can explosively release key cytokines such as interleukin (IL)-4 and interferon (IFN)-γ upon TCR engagement and may regulate a variety of infectious and autoimmune conditions. Here, we report the existence of a second subset of CD1d-restricted CD4⁺ T cells that do not express the NK1.1 receptor or the Vα14 TCR. Like the Vα14⁺ NK1.1⁺ T cells, these T cells exhibit a high frequency of autoreactivity to CD1d, use a restricted albeit distinct set of TCR gene families, and contribute to the early burst of IL-4 and IFN-γ induced by intravenous injection of anti-CD3. However, the Vα14⁺ NK1.1⁺ and Vα14⁻ NK1.1⁻ T cells differ markedly in their requirements for self-antigen presentation. Antigen presentation to the Vα14⁺ NK1.1⁺ cells requires endosomal targeting of CD1d through a tail-encoded tyrosine-based motif, whereas antigen presentation to the Vα14⁻ NK1.1⁻ cells does not. These experiments suggest the existence of two phenotypically different subsets of CD1d-restricted T cells that survey self-antigens loaded in distinct cellular compartments.