Flowering time and the natural distribution and dry matter production of Townsville lucerne (Stylosanthes humilis) populations

Abstract

The flowering times of 58 collections of Townsville lucerne from typical sites in northern Australia have been recorded in three spaced plant experiments near Townsville. Thirty-six collections were grown in 1963-64, 15 in 1964-65, and 17 in 1965-66. All the late flowering collections came from sites receiving at least 45 inches annual rainfall. The five collections from south of Rockhampton were all of the early or midseason type and all collections from the far northern areas were late flowering. Partial regression analysis was used to relate the flowering time of a collection to the rainfall (for the five months interval from January to May) and latitude of the collection site. In the first two experiments rainfall, latitude and (latitude)2 all contributed significantly to the regressions, but in the third experiment only rainfall was significant. Correlation coefficients for 1963-64, 1964-65, and 1965-66 were +0.83, +0.97, and +0.93 respectively. A selection was derived from a collection by bulking seed from single spaced plants selected for uniform flowering time and growth habit. The dry matter yields of some collections and selections were compared in two sward experiments near Townsville in 1964-65 and 1965-66. In 1964-65 there were significant yield differences between collections (experiment A, P<0.01) and between selections (experiment B, P<0.001). There were differences in the rates of vegetative growth and differences in the length of growing season, with late flowering types being able to make better growth late in the season when early types were flowering and seeding. Types with erect growth habit had the highest yields and seemed to compete better with sown grasses than the prostrate types. There were no significant yield differences in 1965-66, a very dry year, and the late flowering types failed to set seed.