The distribution of solvent in block copolymer solutions near their order-disorder transitions is examined experimentally, by small-angle neutron scattering (SANS), and theoretically, by the self-consistent mean-field (SCMF) approach. Three lamellar-forming poly(styrene-b-isoprene) diblocks were employed, in toluene, a neutral good solvent, and in cyclohexane, a selective solvent. For a given copolymer concentration, two solutions were prepared, one in protonated and one in perdeuterated solvent, and the scattering profiles compared. For a neutral solvent, one expects a small partitioning of the solvent to the interface between microdomains, to screen unfavorable styrene-isoprene contacts. Such partitioning should be manifest as a difference between the h- and d-solvents in the intensities of the second (and higher order, even) harmonic peaks (i.e., I(2q*), where the lamellar spacing, L, is 2π/q*). This difference is observed experimentally, and is in quantitative agreement with SCMF predictions using literature values for the three interaction parameters. Interestingly, the predicted relative scattering intensity between solutions in h- and d-toluene varies by several orders of magnitude over a small range in copolymer composition or over a small range of relative interaction parameters of the solvent for the two blocks, suggesting that SANS could be used in this way as a very sensitive measure of copolymer composition and/or solvent selectivity. For a selective solvent, one expects partitioning of the solvent between microdomains and a concomitant change in the intensities of the primary (and higher order, odd harmonic) peaks between h- and d-solvents. This effect is seen clearly in cyclohexane, with the partitioning of the solvent into the isoprene domains increasing with decreasing temperature. However, the SCMF calculations can only match the behavior in both solvents by employing polyisoprene-cyclohexane and polystyrene-cyclohexane interaction parameters significantly different from those reported in the literature.