Genome size in Allium:: In quest of reproducible data

Comparing Allium genome size measurements of different authors, we noticed that the estimates for certain species diverge more strongly than one would have expected in view of the methodological advantages of the material. As the matter has theoretical significance for explaining the biological role of genome size variation, we measured, by Feulgen densitometry, 28 species and altogether 57 accessions or cultivars. Flow cytometric measurements supplemented these data. The current hypothesis of a discontinuous and step-wise distribution of DNA amounts in Allium seems questionable, as most of our DNA values did not appear in the corresponding DNA group as proposed previously. On the other hand, we can confirm that there is a significant negative correlation between genome size of a species and its first month of flowering, but only in diploids, or in diploids and polyploids if only the basic genome size (2Cx level) is considered. We compared our results with those of nine other publications. Only 29 of 60 2C values published previously deviate less than 10% from our data, the others deviate more strongly, from 0.44- to 1.44-fold. The more comprehensive data sets of various authors were compared by correlation analysis with our data. Positive and mostly significant correlations were seen in all tests, but nevertheless the degree of incongruence between studies was unsatisfactory in view of the much better intra-laboratory reproducibility of the present data. The present work highlights the need for generally agreed improvements in standardization and preparative procedures of cytophotometric genome size determination. (C) 1999 Annals of Botany Company.