Flames dance along the underside of two large pulsating sheets of water. The flames trapped underneath by the water are dragged along and eventually roll back due to their own heat.

Fire is the original light. Water it’s existential opposite.
These opposites interact without destroying each other. Together they create original and unpredictable ‘flowers’ in a dance of light.

While we’re looking at the growth and death of these ‘flowers’, we understand that an intriguing physical process is taking place between the fire and the water above it.

In front of the Schreierstoren


In this project we look directly at the source of the light rather than at it’s projection.

Fire is the original artificial light. Here light acts almost as a material… we animate the material using normal DMX software.


On a timeline we animate the pressure and duration of the sheet of water. The 10 flames are each individually animated for duration and pressure.

The flames are made with propane and are electronically ignited. Water shot at a polished cone generates a sheet of water 5 mtrs. in diameter.

The action results in a short-lived ‘flower’, (three minutes approx). This can be organised to coincide with the canal boat timetable.

There will be two ‘flowers’, which can evolve side by side but which can also be made to crash into each other during the programme.

The system will be made with automatic cut-offs to be ‘fail safe’.
Outside display hours the machine will be disarmed.

A sheet of water of 5 mtrs. diameter at a height of 6 mtrs. requires a HF pump of at least 7,5 kW. The system will require a 380V supply rated at 40 kW.


This chart shows differing values calculated for a sheet of water of 5mm., 4mm., 3mm., 2mm., 1mm. and 0,5 mm. thickness at the edge of a circle of 5 mtrs. diameter.

The pump we have chosen is considerably over-rated. In practice it may be possible to go higher than the 6 mtr. above water level (5 mtr. above barge), that we are now proposing.



The propane bottles will be stored open on deck (or concealed in an open space on the barge). The system is outdoors with no closed spaces for the gas to accumulate. The bottles will never be stored in the hold of the barge.

Outside these hours the machine will be disarmed.


Two bottles will need to be replaced every day. We will use a total of 120 x 10 kg. bottles.


In the case of high wind the system will cut off. There is an anemometer connected to the system.

(Rain does not inhibit the presentation. In fact rain drops interacting with the water of the ‘flowers’ should give interesting new forms. The flames will be partially shielded from the rain by the sheet of water above).


In the case of no ignition the system will cut off. We use infra-red sensors to detect the presence of flame.

(In fact, since the ‘gas system’ is in the open air, failure to ignite does not present a danger).


The two ‘flowers’ will be built like a flag pole. For maintenance we will be able to rotate the pole 180 degrees to access the flame system and the nozzle without the need of scaffolding.


The rendering shows the piece set up in front of the Schreierstoren.

Last year’s entry was shown in front of the Hortus in Amsterdam. (This is where the title ‘A Thousand Flowers’ came from).

We could consider any other site…


The ‘flowers’ are at least 5 mtrs. in diameter. The water of the the ‘flower’ continues to travel and will eventually break up, making a natural perimeter of 8-10 mtrs. Nobody would normally even want to go inside this perimeter of falling water.

If they were to do so they would get very wet but not get near the flames which are at 6 mtrs. above them and burn upwards. The fire burns only for short periods.


Rigging for the testing set up

The tests used a pilot light instead of electric ignition. The water was controlled by solenoids.