.  
At 18-22 mph (on a Bell), the annular ring shortens in the direction of movement to become an ellipse, coinciding with translation. Above 20 mph, the annular ring disappears, and a large amount of separate, small airflows coalesce to provide an area of ill-defined airflow with a general downward direction.

.  
Above 35 mph, two distinct rotating vortices are formed from directly behind the machine to a long way behind, assuming no outside influence (they are fully developed about 1 rotor diameter behind the mast, and can be sustained for up to 2500 feet). Each vortex starts from where the annular ring would be in the hover, and is relatively calm in the center (in fact, the centre to centre distance between them is just under the rotor diameter, and slightly displaced from the centre towards the retreating

blade). Regard them as large
funnels extending rearward and
downward, getting bigger as
they go. The point is, they can
be used to direct chemical into
the foliage. Ground cushion,
however, can cause them to
separate, because there is
nowhere else for them to go.
There is still a downward flow
as well as the vortices.
Particle size at low heights should be larger to prevent them being sucked into the vortex areas, where they will not be effective – low level operations should plan to use the lower parts of the generated airflow (say for economical application of pesticides, or where penetration is not required). You would spray at higher levels if you want to spray the underside of foliage, as with an orchard, and make full use of the vortices. In this case, particle size would be smaller, but large enough to fall out of the air stream. The denser the foliage, the higher the air stream needs to be.
It has been found that, between 40-80 mph, the swath width can be relatively constant, meaning that you can get the same physical coverage while slowing down to concentrate on more heavily infested areas. Application is inversely proportional to airspeed, so 2 gallons an acre at 60 mph translates to 3 gallons at 40 or
1.5 at 80.
Note that there is a different optimum particle size according to the foliage - a good reference book for your library in this respect is
Concentrated Spray Equipment, Mixtures and Application Methods, by S F Potts (Dorland Books, NJ).

The Procedure
Crop-spraying, like slinging, is very satisfying when you get a good rhythm and an efficient team that keeps you in the air as much as possible. Unlike it, however, you will be operating a heavy machine with unwieldy spray booms (actually just like bouncy missiles, if you ever did that sort of thing) at some speed in confined areas, and the low level manoeuvres will require a lot of co-ordination and forward thinking. For example, you have to continually keep a note of the ground so when you turn round you can pick up from where you left off and remember where you've been already. When lining up, you need to get the speed and heading right, and pick a spot to aim for so you can keep straight, all in the space of a few seconds. There are GPS-based systems that will even turn the booms on and off automatically. There's usually very little wind to help you, either, because of the legal restrictions on wind speeds and the possibilities of Spray Drift (see later), and spraying cotton in Australia is actually done at night. In fact, you can only really count on about 5 hours' productive time during the day, 3 in the morning and 2 in the evening. Anything else is a bonus. The big problem with doing it at those times is that the Sun is always in your eyes from one direction, which may not be the ideal if there is any wind about.