Improvement of an elutriation method for estimation of weed seedbank in the soil

Abstract

A model experiment was conducted to compare hand extraction (elutriation) under running water and elutriation using the device Analysette 3, the two methods for estimation of weed seedbank in the soil. Technical parameters have been assessed for efficient operating of the device. We studied the time required for elutriation of soil samples and the time for collecting, counting and identifying the separated seeds. No significant difference in the effect of the used elutriation method on results of qualitative and quantitative estimation of weed seedbank densities has been found at any of the three locations differing in soil texture (silty loam, loam and clay loam soil). The time necessary for elutriation of soil samples was highly significantly shorter if the device was used, by 35.5 to 42.9% depending on soil texture vs. hand elutriation. The shortest time was assessed for silty loam soil. By contrast, the time needed for selecting, counting and identifying seeds was shortest for clay loam soil. This time was 46 and 92% longer for loam and silty loam soil, respectively. These differences were statistically highly significant. Comparing the seedbank in the soil, a significantly lower number of weed seeds as well as species was found on silty loam soil vs. the two locations with heavier soils. Amaranthus retroflexus was a dominating species at all locations, and on silty loam soil also Chenopodium album. Of a total number of 32 weed species detected in our experiment, 28 were annual and only four perennial (Cirsium arvense, Elytrigia repens, Rumex obtusifolius and Sonchus arvensis).