Fire has long been used to manage Flint Hills prairie in Kansas (central USA). A study was undertaken to assess True Prairie vegetation response to winter and spring burning. Main grasses were big bluestem Andropogon gerardi, little bluestem A.scoparius, indiangrass Sorghastrum nutans, switchgrass Panicum virgatum, Kentucky bluegrass Poa pratensis (non-native) and sideoats grama Bouteloua curtipendula.
In 1928, two sets of five plots (protected from livestock grazing; no details of plot size given in the original paper) were established, one set burned annually and one biennially. Approximate burn dates were: winter (1 December), early spring (20 March) mid-spring (10 April), late-spring (1 May), and unburned (control). Treatments were suspended in 1944 but resumed in 1950, plots being burned annually giving two replications of each treatment.
In 1950, three 44 acre pastures were fenced for a burning-grazing trial, plus an unburned 60 acre pasture for comparison. Burns dates were the same as on the plots. Cattle (moderate stocking rate of 5 acres/animal unit) grazed from about 1 May to 1 October each year.
Plant species basal area was recorded in the plots and pastures using a randomized line transect method. In 1959-1960, two aluminum tubes were installed per plot and soil moisture recorded at intervals through each year.
Late spring burning (which sometimes was just after growth commenced dependant on weather conditions) was identified as the least detrimental burn time in terms of desired native flora. Thus a more detailed comparison of late-spring burning and no burning was made.
The advantages of late-spring burning over not burning were: an increase in big bluestem, and control of Kentucky bluegrass, Japanese brome Bromus japonicus (both non-native) and buckbrush Symphoricarpos orbiculatus.
Disadvantages were: reduced water infiltration rate soil moisture and forage yields, and increases of smooth sumac Rhus glabra (an undesirable shrub, sometimes considered invasive) which was not controlled by burning.