Those two extraction percentages do not sit on the same footing. Seed came out of the Capel windrows by hand, because the second stage of the proposed system, mechanical pickup and threshing, was never run [7]. So 84.2% describes seed that a digger placed in a windrow and people then collected, while 61.3% describes what a suction machine actually delivered into a hopper [11][12]. The 22.9-point difference [2] is the ceiling available to the swathing route, not a yield, and every loss a pickup head and thresher would add comes off it.
The yield finding is the durable part. Potential seed yield did not separate the two November and December cuts from the February harvest, which says the crop had largely finished making seed well before it finished drying [9]. Standing through summer was not adding to the total. Germination on the Dec. 1 seed came in at 98.0% against 97.5% for the February seed [14], so at that date earliness cost nothing measurable in quality. At the earlier date it did: the Nov. 25 seed was lighter, with lower and more variable germination [13], although the seedlings it produced were of broadly similar vigour [15].
Then there is the schedule. Losing 15.8 points of extraction efficiency across six days works out at roughly 2.6 points a day [3], on a crop whose conventional harvest came at least 62 days later [4]. A grower would have to read that window, and the most obvious instrument for reading it is not yet dependable: burr attachment strength predicted loss only in some of the greener, less mature plants [18], and soil moisture, crop biomass, burr burial, soil conditions and seasonal weather are all likely to move the timing as well [19]. The authors call the practical window relatively narrow and variable between cultivars, paddocks and seasons [24]. Because the design leaves cut plants drying in windrows before collection [4], whatever weather falls between the cut and the pickup is a risk the untested second stage would have to absorb [7].
My read: as a stage-one result this deserves the next round of funding, since it retires the assumption that the crop must be bone dry before its seed is worth taking [22]. As a harvest system it is one site, one season and two cutting dates, with the threshing end still hypothetical [5]. The paper's own verdict, technically possible and worthy of further development [23], is at the right altitude.
