Noble gas isotope systematics, particular those of He, have been fundamental in showing that some ocean island basalts (OIB) were sourced from deep mantle plumes. Relationships between He, W, Os, Sr, Nd and Pb isotopes in Hawaiian, Samoan, Galapagos, and Icelandic lavas have been suggested to reflect contributions from less degassed lower mantle sources, and perhaps even materials advected from the core-mantle boundary. This study reviews the noble gas (He-Ne-Ar-Xe) isotope systematics of major OIB suites. Important in this evaluation are considerations of spatial and temporal variations, including the sample media (glass, minerals, hydrothermal gases and fluids) used for analyzing noble gases, as well as the degree of partial melting experienced to produce OIB. Limited availability of gas-rich samples means patchy coverage in definition of OIB Ne, Ar, Xe isotope compositions. Additionally, low-degree partial melting will lead to preferential sampling of more fusible, generally more enriched components which could affect noble gas isotope systematics. OIB with low- (<8RA) to MORB-like 3He/4He (8 ± 2RA) dominantly sample convecting mantle domains and can also contain some relatively undegassed (solar) components. Their range in Sr-Nd-Os-Pb isotope compositions reflect a strongly recycled crustal and/or lithospheric heritage. Intermediate (>10RA) to high-3He/4He (>25RA) OIB (Loihi, Hawaii; Iceland; Fernandina, Galapagos; Ofu, Samoa) sample a reservoir that has been relatively isolated since ~4.5 billion years as shown by noble gases (He, Ne, Ar, Xe), and by W isotopes. However, this reservoir is not pristine and these OIB show evidence for containing depleted and enriched recycled components from Xe isotopes as well as lithophile-siderophile radiogenic isotope systematics. Linking the highest-3He/4He OIB to a putative Focus Zone (FOZO) reservoir is also problematic; FOZO is not a reservoir least affected by recycling of crust, rather a mantle reservoir that contains recycled components, including depleted lithosphere, and that has been oxidized. This distinction in definition means that, while the FOZO reservoir is almost certainly sampled by the deepest, hottest mantle plumes, its status as a primitive reservoir is unsubstantiated. Models of mantle convection can satisfy seismic constraints on Earth's deep mantle and reveal extensive mixing throughout with enriched and depleted lithologies formed by partial melting processes at Earth's surface. Ancient deep isolated reservoirs are likely to be relatively minor present-day mantle features. More remarkable is the predicted abundance of strongly refractory mantle material. Focus on the noble gas, radiogenic and stable isotope attributes of such refractory reservoirs are required for a fuller understanding of mantle processes.