Field and laboratory experiments were conducted with unicorn-plant to determine conditions required for germination and to measure seed production. Field-grown plants produced an average of 122 pods/plant with an average of 71 seed/pod. The highest percent germination from seed harvested in 1980 occurred following a 2-week prechill treatment of 4 C. Germination increase was greater by removing the seed coat and the membrane enclosing the embryo than by removing the seed coat alone. Germination of seed from all pod compartments was similar. Aqueous extracts of unicorn-plant testa, leaf, stem, root, and exocarp were inhibitory in petri dish bioassays to cotton radicle growth. Extracts of stem, root, and exocarp were inhibitory to wheat radicle growth, and extracts of endocarp, leaf, and exocarp were inhibitory to unicorn-plant radicle growth. Seed buried in the field 10 cm deep for 1 to 8 months showed increased germination over time. Germination was lower when seed were stored at 4 C for 1 to 8 months in a soil having 25% (v/v) water.

Soil-water profiles of cotton, several weeds, and bare soil were compared in research conducted in 1989 and 1990 near Perkins, OK. Neutron probe access tubes were installed prior to plant establishment to facilitate nondestructive volumetric water content determinations at selected depths throughout the season. Planting was in a double-circle pattern concentric around each neutron probe access tube. Phenological and soil-water data were collected weekly. As plants began to senesce, aboveground biomass yield data were collected. In 1989, rainfall frequently replenished soil-water in the upper soil profile. The plants used the recharge water to varying degrees, particularly late in the season. Common cocklebur and johnsongrass extracted soil-water from greater depths than cotton and velvetleaf throughout the season. In 1990, the largest differences detected between bare soil and weeds early in the season were measured for common cocklebur, devil's-claw, and silverleaf nightshade. Later in the season, however, there was little difference in soil-water content in the upper soil profiles for all species. Johnsongrass emerged later than the other species in 1990 which affected its relative soil-water extraction.

The duration and intensity of unicorn-plant interference on lint yield of cotton were evaluated in the field. Random densities of 5.5 ± 1.1 unicorn-plant/m2 reduced lint yield by 41 kg/ha or about 5% for each week that unicorn-plant was present. Interference by 4, 8, and 12 weeds/10 m row decreased yield by 22, 49, and 56 kg/ha, respectively, for each week of weed interference. Each 1 kg/ha of unicorn-plant dry weight reduced lint yield by 0.26 kg/ha. Linear regression of weed dry weight could be used to predict cotton lint yield changes regardless of duration or intensity of weed interference.

The distance-of-influence of an individual devil's-claw plant to cotton leaf, stem, boll (reproductive parts), and combined aboveground plant parts was determined in two field experiments. The distance-of-influence could not be detected for the first 6 weeks after emergence; by 9 to 12 weeks, it extended up to 25 cm or more; and by the end of the season, up to 50 cm on each side of the weed. Cotton leaf and stem weights were less sensitive than cotton boll weights for measuring distance-of-influence from devil's-claw. At maturity, cotton boll weight was reduced 62 and 51% in 1983 and 45 and 29% in 1984 for sampling intervals of 0 to 25 and 25 to 50 cm, respectively. Interference from cotton reduced devil's-claw stem weight, seed capsule weight, and whole plant biomass by 6 weeks after emergence and reduced leaf biomass by 9 weeks in 1983. All except stem biomass were affected in 1984. Distance-of-influence and weed-density interference studies predicted lint yield loss similarly.

Soil water from plots containing cotton, devil's-claw, cotton with devil's-claw, and bare soil was measured throughout the growing season using a neutron probe and related to weed interference with the crop. Volumetric water content throughout the soil profile to a depth of 180 cm did not differ among treatments before the 5th or 6th week after cotton emergence. Greater water depletion occurred early in the season in plots containing devil's-claw which corresponded to a period of rapid weed growth. In plots containing only cotton, the largest reduction in water content occurred later in the season during peak bloom and early boll formation. Soil water content at depths greater than 105 cm remained unchanged in all plots throughout the season. Interference from devil's-claw reduced cotton lint yield 96% in 1986 and 46% in 1987. Higher rainfall and reduced weed populations in 1987 reduced the impact of weed interference on cotton lint yield.

Full-season interference from unicorn-plant [Proboscidea louisianica (Mill.) Thell. # PROLO] at densities ranging from 0 to 32 weeds/10 m row was measured on weed and cotton (Gossypium hirsutum L. ‘Westburn M′) parameters in three Oklahoma environments. Linear regression models using log10 transformations were superior to linear and comparable to quadratic models in estimating the impact of inter- and intraspecific competition. Weed measurements on an individual plant basis were not reliable indicators of interference. Weed dry biomass, ground cover, and seed capsule production/plot generally increased with increasing weed densities. In the three environments, densities of two, four, and eight weeds/10 m row initially reduced cotton plant height. Maximum height reductions averaged 20, 28, and 43% in the three environments but did not detrimentally affect the mechanical harvest of cotton. As unicorn-plant density (expressed in log10 units) doubled (within a range of 1 to 32 plants/10 m of row), lint yield reductions ranged from 84 to 146 kg/ha. Maximum lint yield losses averaged 59, 65, and 74% in the three environments. Transforming lint yield/plot to a percentage of that from weed-free plots resulted in a single quadratic equation applicable over environments. Unicorn-plant at the highest weed density reduced fiber fineness, uniformity, and length, but not fiber strength.

Field experiments conducted in 1991, 1992, and 1993 evaluated Palmer amaranth and devil's-claw control and cotton injury with pyrithiobac applied PPI, PRE, or POST. Pyrithiobac at 36 or 71 g ae/ha applied PPI, PRE, or POST did not injure cotton. Pyrithiobac at 140 g/ha applied PPI or PRE injured cotton 9 to 11% 6 wk after treatment. Cotton recovered and no injury was observed 12 wk after treatment. Pyrithiobac applied PPI and PRE at 71 g/ha controlled Palmer amaranth at least 97% 6 wk after treatment. Palmer amaranth control with pyrithiobac applied POST was more variable and influenced by environmental conditions. Palmer amaranth control with 71 g/ha of pyrithiobac exceeded that with 36 g/ha. Devil's-claw control with pyrithiobac was better with POST applications than PPI or PRE applications. Pyrithiobac applied POST at 140 g/ha controlled devil's-claw 83–97%. These studies indicate that pyrithiobac can effectively control Palmer amaranth and devil's-claw in cotton on the Texas Southern High Plains when applied at appropriate rates and timings.

Devil's-claw and smellmelon are summer annual broadleaf weed problems in the southwest cotton production region. Field studies conducted in 1993 and 1994 evaluated the efficacy of pyrithiobac for postemergence (POST) control of devil's-claw and smellmelon compared with soil-applied fluometuron or prometryn. Pyrithiobac at 0.07 and 0.14 kg ai/ha early POST and 0.14 kg ai/ha mid-POST controlled devil's-claw more than 84% at the late rating. Devil's-claw control with mid-POST applications of pyrithiobac at 0.07 kg ai/ha was reduced when compared with other rates and application timings. Smellmelon control with pyrithiobac ranged from 57 to 84% and, except when applied at 0.07 kg ai/ha early POST, was equivalent to or greater than either fluometuron or prometryn. Crop injury with pyrithiobac treatments 16 d after treatment was 5 to 13%, but 33 d later, was no more than 3%.