Understanding how magma is stored within the crust, and what processes it might undergo, is of paramount concern to Earth scientists. Whilst advances in analytical techniques have revealed increasingly complex pictures of magma storage at arc and mid-ocean ridge settings, we still know relatively little about continental rift volcanism. The Main Ethiopian Rift (MER) is part of the larger East African Rift system, the archetypal example of continental rifting on our planet. Comprehensive geochemical datasets, including new whole-rock, glass (matrix glass and melt inclusions), and mineral analyses, have been generated for several geologically-young volcanic sites across and along the MER. Quaternary eruptive products are typically bimodal. Whole-rock and glass compositions can be modelled effectively by protracted fractional crystallisation of the alkali basalt to produce pantelleritic rhyolites. Despite this seemingly simple geochemical signature, alkali feldspars show trace element (Ba) heterogeneity suggestive of more complex processing. Many are antecrysts, crystallised from a more primitive melt, before being efficiently segregated due to the inherently low viscosity of peralkaline liquids. They are then later entrained from a ‘crystal-mush’ region into more evolved residual melts during large, explosive eruptions. Melt volatile contents are also inconsistent with a simple fractional crystallisation model. Basalts stored both on and off axis are saturated in H2O and CO2 at depths up to ~30 km in the crust. Much of this volcanic CO2 is lost through diffuse surface degassing, measured CO2 fluxes require the intrusion up to 0.14 km3 of basalt beneath the MER each year. On axis a change in S, Cl, and F solubility is recorded over the compositional gap, likely due to a drop-in pressure, as magmas ascend to a shallower storage region (~5-8 km), and the rapid increase in SiO2. The build-up of a pre-eruptive exsolved volatile phase has implications for ore formation and volcanic hazard assessment. The loss of halogens from MER melts likely precludes them from being a viable source of economic REE mineralisation, meanwhile low fluid/melt partition coefficients means that S yields of peralkaline eruptions are predicted to be much higher than their metaluminous counterparts. The state of a magma, in terms of composition or phase, can dramatically influence the geophysical signals observed at the surface. For example, the enhanced compressibility of an exsolved volatile phase will dampen geodetic monitoring signals. Magnetotelluric (MT) surveys are commonly used in Ethiopia, however, have sometimes produced results that contradict other geophysical methods. Forward models of electrical resistivity beneath the MER, informed by geochemical constraints, can help improve MT interpretation. The absence of a conductive body, synonymous with the presence of melt, beneath Aluto volcano can be reconciled with other geophysical evidence if the magma is highly crystalline. The presence of conductive anomalies along and across axis likely reflects the presence of longer-lived, more melt-rich systems, highlighting spatial variation in magmatism common to rift settings. The poor spatial resolution of the MT technique limits its use for volcanic monitoring in the MER. Only the influx of large volumes of melt, or rejuvenation of a crystalline storage system, perhaps prior to larger eruptions, may be detectable on repeat measurements.