By producing thermal fluctuations with a mandohe and helium fluctuations with chimneys attached to the grid bars, the mixing of temperature and helium fluctuations as well as the decay of temperature and helium variance and their flux is investigated in decaying grid-generated turbulence. The helium, temperature and velocity fluctua- tions were measured with a modified ‘Way-Libby’ interference probe (Way & Libby 1970, 1971). It is shown that, as for temperature variance, the helium-variance decay rate is a function the ratio of the helium length scale to the velocity length scale. It is also shown that the decay of the cross-correlation between temperature and helium fluctuations is slow if both scalars are introduced close to the grid, but rapid if each scalar is introduced at a different distance from the grid, and hence at different scales. The results corroborate those of the inference method of Warhaft (1981), which is extended here to examine other cases. A particularly unexpected finding is that under certain circumstances the two-scalar cross-correlation may actually increase with distance from the grid, although the scalar covariance decreases. The return to isotropy of helium flux and temperature flux is also investigated and is shown to be slow if the scalar flux is produced near the grid bars, but faster if the flux is produced further downstream. For all the measurements helium and temperature were passive additives.