Recurrent phenotypic selection for low grasshopper food preference in rangeland alfalfa

Abstract

Grasshopper [Melanoplus spp. and Cannula Pellucida (Scudder)] feeding is an important factor that can prevent establishment and reduce yields of alfalfa [Medicago sativa subsp. X varia (Martyn) Arcang.] interseeded into semiarid rangelands of the northern Great Plains. Objectives of this study were to determine narrow-sense heritability estimates of grasshopper preference for an alfalfa population and to develop low-preference alfalfa germplasm adapted to rangeland use. The base population was derived from 5 cultivars and an experimental strain, all of M. sativa L. subsp. sativa X M. sativa subsp. falcata (L.) Arcang. parentage, that had been developed for rangeland use in the northern Great Plains. Except for cycle 4 which was conducted in a greenhouse, each cycle of recurrent phenotypic selection involved transferring replicated half-sib families of 6- to 8-week-old plants in flats to a field site with a heavy infestation of grasshoppers. Half-sib families were scored for defoliation when the entire population was at least 50% defoliated. Narrow-sense heritability estimates of defoliation under greenhouse conditions for 31 half-sib families and their respective parents ranged from 45 to 58%, depending on how defoliation from the multi-specks grasshopper populations was measured. Alfalfa populations produced from cycles 1, 3, and 5 of recurrent phenotypic selection were evaluated simultaneously under greenhouse conditions with a common grasshopper population and rated for defoliation on a scale from 1-5 where 1 = 0-20% and 5 = 81-100% defoliation. Mean defoliation decreased significantly (P<0.05) from 3.83 to 3.25 from cycle 1 to cycle 3, but the small decrease from 3.25 to 3.15 from cycle 3 to cycle 5 was not significant. The lack of progress from cycle 3 to cycle 5 was attributed to a major change in species composition of the grasshopper populations used in the selection process.