I would suggest that this is best done by an indirect
method such as diffraction or interference.
You will presumably know the frequency of
your ultrasonic pulses and so if these are transmitted through the liquid and allied to diffract
through a slit their wavelength can be found. Hence the velocity of the waves can be
determined from:
Velocity = frequency x wavelength
There is a little problem in
knowing the size of slit to use but the approximate order of size could be estimated by
calculating the wavelength of the waves from their velocity in air or free space (300 000
km/s) and then making some assumptions as to whether you think that they will go faster or
slower in the liquid under investigation.