Every time the hot steam was bubbling up, the water and plant started to boil.
The water/plant mixture probably didn't 'boil' in a physical sense, that is undergoing a phase change from liquid to gaseous.
The 'boiling' or more like 'agitation' of the mixture was caused by the introduction of a water vapor. More on that later...
SWIM used to flask, one with steam and one with plant. It was easy enough to get the oliebath in the steam flask to work, but it was pretty hard to keep the plant flask under the 100c.
Why would you want to keep the flask with plant/water mixture below 100°C? The solution has to be at its boiling point, otherwise the introduced steam (unless superheated) will condense when introduced to the mixture instead of passing it, and carrying over the ephedrine. That explains why none was found in the reception vessel.
The flask will also lose heat at its surface, causing water to condense, increasing the solution volume. This has to be counteracted by _heating_ the plant/water vessel, in order to keep its liquid volume constant.
To make things worse, the boiling point of a concentrated aqueous solution at atmospheric pressure is _not_ 100°C, but can be significantly higher. For example, the bp of an aqueous 20% (by weight) NaOH solution is 108°. Soluble components introduced with the plant matter would increase the bp even more.
So heating that flask in definitely in order, unless superheated steam is used. For a 1l rb flask SWIN uses a heating mantle set to 100W power with good results.
For your setup, the correct amount of heating can be determined to observing the volume level in the water/plant mixture flask - if it increases, more heating is required, if it decreases lower the heat.