In order to get a gradual change in the density of a solution, you can connect two identically sized containers holding pure and salt water respectively with a pipe, and stir the pure water while pouring the two water types into the water tank near the base, as shown in the diagram. Of course, you can pour the heavier solution in first, but then you have to use a float to move the spout up with the water level as you pour the lighter solution.
When using this technique, it is assumed that the pressure at the base of each tank (where the pipe connects) is statically balanced. So the thicker and shorter the pipe connecting the two tanks is the better. However, if the pipe connecting the two tanks is straight, the pure water will flow backwards along the upper part of the pipe, so the pipe should be a U-shape. Also, if the pouring speed is too fast, the two fluids will mix. You can prevent this by placing a rough sponge or something similar near the pouring spout to absorb the moving water's energy. (It is best to use a sponge with a large surface area. Putting a piece of sponge in the pouring spout will have no effect.)
In order to measure the strength stratification (density gradient), it is easiest to take a sample of each layer and measure the concentration of salt. In both the internal gravity wave and quasi-biannual oscillation experiments, we measured the density of last of the salt solution in the mixing tank, and calculated the average density gradient from that.
However, in order to determine the distribution of stratification strength in certain specific locations, you have to compare the difference in specific gravity of nearby layers. What we did then was to shine a laser horizontally into the stratified fluid, and calculate the distribution of density with the angle of refraction of the laser beam. It is difficult using this technique to get an accurate absolute figure for the density gradient, but it makes it easy to get an idea of the relative distribution over a wide area without direct contact with the working fluid.
When actually measuring, it is necessary to have a standard beam for comparison, so in the experiment, we used a cylindrical lens (or glass rod) to get a sheet of light, and then placed one mirror in front of, and one behind the water tank, and measured the difference in angle of the light reflected back off the two mirrors.