Because we may wish to accurately squirt small amounts of liquid (like a just small number of drops). We might want to squirt small amounts of herbicide into specific plants. We may wish to annoy animal pests and discourage them from feasting in the veggie patch. We may wish to squirt human trespassers with unsanitary liquids to discourage them from remaining on the property.
If aiming requires two degrees of freedom, it can be done with two servos connected in series at 90°.
These Dynamixel digital servos are excellent - though quite expensive.
Actually, there is a better way - see https://www.youtube.com/watch?v=97t7Xj_iBv0&t=609s at 8m18s for a satellite thruster mount flexible machine, but I don't think they will give me one of those.
Alternatively, Trosson robotics sells this amazing two-axis servo mount. But it is very expensive (if you have to ask, you can't afford it).
This is harder.
Several possibilities come to mind, but the basic design
water reservoir --> pump --> flexible tube --> solenoid --> nozzle --> spray.
has great appeal. Mostly because I have done similar things before and if nothing else, resembles familiar irrigation systems. The pump would be expected to deliver a reasonably constant pressure. Maybe the voltage to the pump could be modified if the pressure needed to be changed.
Most nozzles used for herbicides are designed to create a spray or mist to distribute liquid around a large area more or less uniformly. This is needed for general application. However very specific application requires accuracy from a distance of 600mm or more. A solid stream is required, and hence a 'solid stream' nozzle is required (it took a while to realise that this was the term to google).
Turning the stream of liquid on and off quickly, and consistently is critical. These solenoids are used for squirting single drops of liquid, so should be fast enough for a controlled stream.
It's a pump. It supplies flow/pressure. It needs fittings to connect it to a flexible hose. Aliexpress is your friend (so if forward planning because the products take a long time to arrive).
Assembly for the solenoid and nozzle mounted on a servo for single-degree-of-axis aiming.
The squirt gun assembly.
The chemical tank (a used drinking water container that we used when our mains water got interrupted).
Hopefully most of this is calibration. Whether the parabolic shape of the stream has to be taken into account will have to be determined by experiment. If the parabolic stream has to be modelled, then so too even light breezes will have to be taken into account.