A PROFICIENCY TESTING PROGRAM FOR THE AGRICULTURE LABORATORY INDUSTRYRESULTS OF THE 1994 PROGRAM

Citation
Ro. Miller et al., A PROFICIENCY TESTING PROGRAM FOR THE AGRICULTURE LABORATORY INDUSTRYRESULTS OF THE 1994 PROGRAM, Communications in soil science and plant analysis, 27(3-4), 1996, pp. 451-461
Citations number
11
Categorie Soggetti
Agriculture Soil Science","Plant Sciences","Chemistry Analytical
ISSN journal
00103624
Volume
27
Issue
3-4
Year of publication
1996
Pages
451 - 461
Database
ISI
SICI code
0010-3624(1996)27:3-4<451:APTPFT>2.0.ZU;2-Q
Abstract
The accuracy of soil and plant analytical results are occasionally cal led into question by laboratory clientele. Although laboratories gener ally conduct internal quality assurance procedures, there are few exte rnal performance testing programs for the industry, In the spring of 1 994, a proficiency testing program was initiated for soil and plant sa mples for agricultural laboratories in the Western United States to pr ovide an external quality control for the laboratory industry. The pro gram involves the quarterly exchange of soil and plant samples on whic h soil salinity, soil fertility, and plant nutrition analyses are cond ucted. For the 1994 program, 103 laboratories from 23 states and Canad ian provinces participated. Overall, soil salinity and pH analyses wer e highly reproducible across soils and laboratories through the year. Variability in soil nitrate results were lower than noted in past inte rlaboratory exchanges independent of soil sample. Results for extracta ble soil bicarbonate phosphorus were variable independent of soil and indicate a significantly positive skewed distribution. Plant nitrogen (N) and phosphorus (P) results were consistent across samples, laborat ories, and methods. Variability in plant nitrate increased with:decrea sing tissue concentrations. Overall accuracy and precision of reported results, based on the use of a NIST certified reference botancial sam ple, were excellent for N, P, potassium (K), calcium (Ca), and copper (Cu). Generally for any given analysis the results of approximately 10 % of the laboratories exceed two standard deviations from the mean. Ov erall, significant improvement was noted in the laboratory industry pr oficiency through the course of the year for the program.