Single point measurement turbulence cannot distinguish variations in space and time. We employ an ensemble of one-and two-point measurements in the solar wind to estimate the space-time correlation function in the comoving plasma frame. The method is illustrated using near Earth spacecraft observations, employing ACE, Geotail, IMP-8, and Wind datasets. New results include an evaluation of both correlation time and correlation length from a single method, and a new assessment of the accuracy of the familiar frozen-in flow approximation. This novel view of the spacetime structure of turbulence may prove essential in exploratory space missions such as Solar Probe Plus and Solar Orbiter for which the frozen-in flow hypothesis may not be a useful approximation.Introduction. The solar wind provides a natural laboratory for fundamental study of plasma turbulence in parameters ranges difficult to achieve in the laboratory (e.g., [1]), and often accessible only through remote sensing [2,3]. Fluctuations in space and time are characteristic of turbulence and each have significant influence in space and astrophysical contexts. The two point correlation functions and associated spectra [4,5] are essential turbulence diagnostics in applications such as plasma heating [6], solar wind acceleration [7], the scattering and transport of energetic particles [8,9], magnetic field connectivity [10,11] and geospace prediction (space weather) [12]. Time correlation enters these applications as well, but in somewhat different ways [9,11], as is also well established in turbulence theory [13]. However, for most in situ interplanetary observations, made by a single spacecraft, there is an almost complete ambiguity between spatial structure and temporal structure. For fast flows, the Taylor frozen-in flow approximation [14] (or, Taylor hypothesis) provides an estimate of purely spatial statistics, with quantifiable limitations on accuracy [15][16][17][18][19]. The time correlation and the space correlation have not been previously treated on the same footing in interplanetary datasets, as far as we are aware. Here we employ a multispacecraft methodology to present a novel view of the space-time correlation of interplanetary turbulence. We show how numerous one-and two-spacecraft measurements may be combined, without use of the Taylor hypothesis, to determine correlation functions in which length and time separations are treated as independent. This method can provide a wealth of information about turbulence structure and dynamics, and may become a valuable tool for interpreting present and future interplanetary datasets.Background. We consider space-time structure of fluctuations in interplanetary space. For simplicity we dis-