Proceedings of the XLVI Italian Society of Agricultural Genetics - SIGA Annual Congress

Giardini Naxos, Italy - 18/21 September, 2002

ISBN 88-900622-3-1

Poster Abstract - 4.02

STATISTICAL MULTIVARIATE ANALYSIS TO CHARACTERISE EMMER (TRITICUM DICOCCUM SCHRANK) GENOTYPES FROM GARFAGNANA AREA

PILERI L., LACERENZA N.G., CHISCI G., BENEDETTELLI S.

Department of Agronomy and Land Management, University of Florence

emmer, discriminant analysis, multiple regression

Six emmer (Triticum dicoccum Schrank) populations collected from localities representative of Garfagnana crop cultivation farms, were cultivated in five different localities with altitudes ranged from 600 m a.s.l. to 980 m a.s.l. on both Serchio river sides, for three years (1995-1998).

The adopted experimental design was complete randomised block with three replications; the elementary plot had a surface of 10 m². The morphological characters were recorded from eight plants randomly chosen within each plot, whereas the variables: weed influence, plant density and production were recorded for each plot.

The data were statistically analysed applying the variance mixed effect model with the accession considered with fixed effect factor and the localities and the year with random effect. To estimate the phenotypic variability between and within population and how this variability can be modified by environmental factors, the data concerning to each locality and to each year have been utilised to calculate the discriminant function. The path analysis was utilised to find connection between yield production and both morphological and environmental factors. This analysis was carried out also considering the type of awnedness (awened, half-awned and awnless), observed in the Garfagnana emmer accession. The dendrogram showed the presence of two main group constituted one by the half-awned and the other by the awnless types, without concerning the belonging to a particular accession.

In conclusion the results of discriminant analysis together with the results of path analysis allow to identify morphological and production variable associations, useful to establish particular selection index useful to be utilised in breeding program of this species.