The solar wind is a structured and complex system, in which the fields vary strongly over a wide range of spatial and temporal scales. As an example, the turbulent activity in the wind affects the evolution in the heliosphere of the integral turbulent scale or correlation length [{\lambda}], usually associated with the breakpoint in the turbulent-energy spectrum that separates the inertial range from the injection range. This large variability of the fields demands a statistical description of the solar wind. In this work, we study the probability distribution function (PDF) of the magnetic autocorrelation lengths observed in the solar wind at different distances from the Sun. We use observations from Helios, ACE, and Ulysses spacecraft. We distinguish between the usual solar wind and one of its transient components (Interplanetary Coronal Mass Ejections, ICMEs), and study also solar wind samples with low and high proton beta [\beta_p ]. We find that in the last 3 regimes the PDF of {\lambda} is a log-normal function, consistent with the multiplicative and non-linear processes that take place in the solar wind, the initial {\lambda} (before the Alfv\'enic point) being larger in ICMEs.

We analyze the evolution of the interplanetary magnetic field spatial structure by examining the inner heliospheric autocorrelation function, using Helios 1 and Helios 2 "in situ" observations. We focus on the evolution of the integral length scale (\lambda) anisotropy associated with the turbulent magnetic fluctuations, with respect to the aging of fluid parcels traveling away from the Sun, and according to whether the measured \lambda is principally parallel (\lambda_parallel) or perpendicular (\lambda_perp) to the direction of a suitably defined local ensemble average magnetic field B0. We analyze a set of 1065 24-hour long intervals (covering full missions). For each interval, we compute the magnetic autocorrelation function, using classical single-spacecraft techniques, and estimate \lambda with help of two different proxies for both Helios datasets. We find that close to the Sun, \lambda_parallel < \lambda_perp. This supports a slab-like spectral model, where the population of fluctuations having wavevector k parallel to B0 is much larger than the one with k-vector perpendicular. A population favoring perpendicular k-vectors would be considered quasi-two dimensional (2D). Moving towards 1 AU, we find a progressive isotropization of \lambda and a trend to reach an inverted abundance, consistent with the well-known result at 1 AU that \lambda_parallel > \lambda_perp, usually interpreted as a dominant quasi-2D picture over the slab picture. Thus, our results are consistent with driving modes having wavevectors parallel to B0 near Sun, and a progressive dynamical spectral transfer of energy to modes with perpendicular wavevectors as the solar wind parcels age while moving from the Sun to 1 AU.