H8nd sep8r8tion of roots of 2 or more plants species from soil cores is a tedious 8nd labo~intensive task. Our objective was to determine whether neu-infrared refierturce spectroscopy (NIRS) could be employed to estimate root biomass proportions in bbmry mixture8 of 8lfaifa (IlYe&ugo sativu L.) with each of 4 grasses. Grasses chosen for experimentation were crested wheatgrass (Agropyron cr&t&m L.), intermedi8te wbe8tgr8ss [Tltbropynun Mefure&uu (Host) Bukwortb & DR. Dewey& an intergeneric

more »

... [E@rigia repens (L.) Nevski X Pseudimegneria spicata (Pumh) Love], urd Russiur wildrye [Psathyrostachys junea (Fiscb.) New&i]. In tbe first experiment, roots from single-species field plots were washed from soil cores, dried, ground, and mechanically mixed in preselected 8lfaKa-grass ratios in which the percentage of grass varied from 0 to 100. Equ8tions to measure the proportion of l lf8lfa or grass were developed from nar btfr8red reflectance data using 84 r8ndomiy selected sample8 In the second experiment, the 5 plant species were grown in greenhouse pots in pure stands urd in birmry mixtures that included all combinations of the gr8sses. Root systems were seprr8ted while 8tt8cbed to the topgrowth, dried, 8nd ground. Tissues from single species tratments were mixed md calibration eqtmtions developed from these mixture8 were used to estimate the proportion of alf8if8 md the proportion of grass in samples. Sunpies contained either one type of root or 8 mixture of roots in proportions sbnil8r to those th8t occurred n8tur8lly in the pots. Coefficients of determination (rZ) between tbe emtimrted and the l cttml root mass r8tlos ranged from 0.92 to 0.99. Determination of tbe proportion of grass in the samples was more 8ccunte and precise than determbmtlon of the proportion of 8lf8If8. After the approprirte crlibrrtion equations have been developed, NlRS is more efficient than hand separrtion for estimrting rlfaifa-grass root mass ratios. The utility of the techniques can be increased by developing equations th8t encompass more complex mixtures and 8 wider range of environment8l circumstances. The relationships among plants in mixed species swards are of concern to ecologists, physiologists, agronomists, and range scientists. Many investigations involve the quantification of plant densities and top growth biomass ratios of the components of species mixtures. Similar measurements of root biomass often are not attempted because of the difficulty of correctly identifying and separating roots of different species. Since the use of near infrared reflectance spectroscopy (NIRS) to measure forage quality was first described (Norris et al. 1976) , the technique has been employed to determine the amounts and