I am on very spongey ground here(thinking without any substantial research nor prior knowledge) but as the 'stream' is rotating about the cone with both circumferential and 'axial'(for lack of a better word) velocity and since velocity is a vector thingie(has magnitude and direction) and the circumferential vector is tangent and constantly being altered by the cone, there must be some interaction between the two vectors. We know the cone is causing the tangential vector to increase in magnitude as the stream moves towards the smaller end. Since that vector is increasing in magnitude
perhaps a similar effect occurs with the 'axial' vector.
However!
Having now just done minimal 'research' I see things a bit simpler.
Horizontal cyclones work by virtue of a secondary air flow which counteracts drag of the stream against the inside wall of the cone and thus allows it to work without dependency upon gravity. The secondary air flow directs the 'debris' to the collector.
A vertical cyclone like the Dust Deputy gets its initial 'downward' velocity by being directed so at the inlet. Gravity continues to exert its pull on the debris as it spends time within the cyclone(and after departing into a collection container). As the debris falls toward the smaller end, it's velocity around the wall increases, but the cone also tends to support the debris. Meantime the air separates from the debris, and travels directly down towards the air outlet tube. This air motion may help move the debris towards the bottom. The air spins less and accelerates towards the outlet thus moving down faster than the debris so it causes the downward velocity of the debris to increase.
Like I said initially, shakey ground!:D
Now just below the air outlet, the air must reverse direction!
I have not a clue what effect that has on the debris stream.
