The paper was right

We started this experiment as a bit of a student project, in a light-hearted frame of mind, but we couldn't get it right for years.

We continually improved our apparatus until we had almost the same conditions as those described in Plum and McEwan's paper, but had no success. Just as we started thinking that if we tried this hard and still couldn't get it right, then there must be something wrong with the original paper, the reversal of the current happened!

During those five years of failures, many students came and went without ever seeing the result of this experiment.

Muddy water is the saviour

The key to success came from a most unusual direction.

Since we could not get it right despite numerous attempts, we decided to use the experiment apparatus for something else. After the Quasi-Biannual Oscillation experiment ended in failure, we poured muddy water on the stratified fluid for a new experiment, and created an internal wave, and a reverse current, although imperfect, emerged!

Encouraged by this result, we made another attempt at the experiment with exactly same parameters, but it still didn't work. The only difference was that we weren't using muddy water, so in desperation, we poured muddy water again, and again the current appeared.

What exactly was going on? Was there something particularly good about using muddy water?

Fight on with Regain (the health drink)

It would appear that the fact that we added water was more important than the addition of mud.

When making the stratified layers in the preparatory stage of the experiment, you pour the freshwater into the tank first, however controlling the process is quite difficult. Sometimes the layers close to the surface don't form properly. Further, even if the preparation goes well, in this experiment you create waves near the surface, so if you continue the experiment over a long period of time, the water near the surface mixes and the stratified layers crumble. It is therefore likely that the stratified layers near the water's surface had already crumbled. By adding water, we inadvertently recreated those stratified layers.

Thus by adding some more water before starting the experiment as some vitamins for the stratified layers, we increased our success rate immensely.

After that, we developed a way of measuring the strength of the stratification, and discovered that the stratified layers to a depth of 2cm from the water surface were crumbling, so instead of adding water, we removed the crumbled layers (by throwing out the water of the upper layers), and finally came up with a repeatedly successful experiment.