We take it to the end of a breakwater . It's a big one which goes out about  a kilometre from shore. The bay is quiet, the only boats around are cargo ships at anchor miles away. The sea is calm - that's important cuz we don't want too much background noise from waves. The day is sunny and bright - not ideal for our purposes. Sharks arre said to prefer dull, overcast days before they'll come near the surface. Early morning is supposed to be a better time as well.
 

It doesn't look much - just a PVC elbow joint with end caps which have been cut to take I/O cables at one end and the acoustic surface at the other. The sound projector is a 0.5mm thick aluminum plate which has a piezo crystal acoustic transducer bonded to its inner surface. Polycarbonate plastic has better acoustic matching properties in water but we couldn't find any in the local hardware shop. There are aluminum brackets on the outside to which a support rope can be attached. It's important that the signal cable does not take any strain. Because it's hollow a ballast weight must be attached to the clamps.
 

We set up the control box & battery supply on the rocks and run the cable down to the water. Then we just plonk the transducer in, switch on and wait to see what happens. There are problems. You can't stop the the thing from twisting so you have no way of controlling which way the acoustic surface is facing. Most of the time it's facing the rocks which won't help fish to hear it.
 
 

What we need to do is trail it from a boat in deep water. But because the object is to attract active, predatory fish like sharks there's NO WAY we're gonna try this from the good ship Thylacoleo. Sailing dinghies are way too small and unstable to do an experiment like this safely. Absolutely, positively the last thing you'd want is to finish up in the water with Mr. Shark.
 

RESULTS?
So what happened? We let it run for about half an hour. It had an effect on small fish - a lot of them seemed to gather in the water near the transducer. You can't really see them in the JPEG with all the reflections. But of larger fish there was no sign. Does that mean the  experiment flunked? No, not really. At this stage we have no idea what sorts of sounds a shark would find interesting, so that aspect will take a lot of experimenting. The Vocaliser can be programmed with all sorts of audio files however. Besides which we don't actually know if even one shark was anywhere near this spot. The transducer is a first attempt : its BW will not be optimal, the power output is probably inadequate and we can't control its attitude in the water. There's obviously a long way to go with this technology so we'd have to say the results are inconclusive right now.

NOTE #1 : I was reminded by J.B. that at the low freqs we're using the small acoustic plate will be effectively omni-directional anyway. Thanks JB!

EXPT #1 : Measure the freq. response of the xtal acoustic element. Should have down this first. Also some notes on the Lloyd Mirror effect.
EXPT #2 : Assessing the feasibility of a magnetorestrictive element as an acoustic transducer.

REFERENCES
Applied Underwater Acoustics by D.G. Tucker & B.K. Gazey, Pergamon Press 1977.
How to Build & Use Low-Cost Hydrophones by Frank Watlington, TAB Books 1979.