Juno Microwave Radiometer (MWR) observations of Jupiter’s mid-latitudes reveal a strong correlation between brightness temperature contrasts and zonal winds, confirming that the banded structure extends throughout the troposphere. However, the microwave brightness gradient is observed to change sign with depth: the belts are microwave-bright in the p<5 bar range and microwave-dark in the p>10 bar range. The transition level (which we call the jovicline) is evident in the MWR 11.5 cm channel, which samples the 5-14 bar range when using the limb-darkening at all emission angles. The transition is located between 4 and 10 bars, and implies that belts change with depth from being NH3-depleted to NH3-enriched, or from physically-warm to physically-cool, or more likely a combination of both. The change in character occurs near the statically stable layer associated with water condensation. The implications of the transition are discussed in terms of ammonia redistribution via meridional circulation cells with opposing flows above and below the water condensation layer, and in terms of the ‘mushball’ precipitation model, which predicts steeper vertical ammonia gradients in the belts versus the zones. We show via the moist thermal wind equation that both the temperature and ammonia interpretations can lead to vertical shear on the zonal winds, but the shear is ~50x weaker if only NH3 gradients are considered. Conversely, if MWR observations are associated with kinetic temperature gradients then it would produce zonal winds that increase in strength down to the jovicline, consistent with Galileo probe measurements; then decay slowly at higher pressures.