The number of angiosperm species for which nuclear DNA amount estimates have been made has nearly trebled since the last collected lists of such values were published, and therefore, publication of a more comprehensive list is over due. This paper lists absolute nuclear DNA amounts for 753 angiosperm species. The dats were assembled primarily for reference purposes, and so the species are listed in alphabetical order, as this was felt to be more helpful to cyto- and biochemists whom, it is anticipated, will be among its major users. The paper also reviews aspects of the history, nomenclature, methods, accuracy and problems of nuclear DNA estimation in angiosperms. No attempt is made to reconsider those aspects of nuclear DNA estimation which have been fully revised previously, although the bibliography of such aspects is given. Instead, the paper is intended as a source of basic information regarding the terminology, practice and limitations of nuclear DNA estimation, especially by Feulgen microdensitometry, as currently practiced.

All else being equal, polyploids are expected to have larger C-values (amount of DNA in the unreplicated gametic nucleus) than their diploid progenitors, increasing in direct proportion with ploidy. This expectation is observed in some polyploid series, especially those newly formed, but there are examples suggesting that C-values in particular polyploids are less than expected. The availability of the Angiosperm DNA C-values database (http:// www.rbgkew.org.uk/cval/homepage.html) has allowed this question to be addressed across a broad range of angiosperms and has revealed striking results deviating from expectation: (i) mean 1C DNA amount did not increase in direct proportion with ploidy, and (ii) mean DNA amount per basic genome (calculated by dividing the 2C value by ploidy) tended to decrease with increasing ploidy. These results suggest that loss of DNA following polyploid formation, or genome downsizing, may be a widespread phenomenon of considerable biological significance. Recent advances in our understanding of the molecular events that take place following polyploid formation together with new data on how DNA amounts can both increase and decrease provide some insights into how genome downsizing may take place. The nature of the evolutionary forces that may be driving DNA loss are also discussed.

We argue that there is a need to maintain the term genome size in a broad sense as a covering term, because it is widely understood, short and phonetically pleasing. Proposals are made for a unified and consensual terminology. In this, 'genome size' should mean the DNA content based on chromosome number x and n, and should be used mainly in a general sense. The necessary distinction of the kinds of genome sizes is made by the adjectives 'monoploid' and the neology 'holoploid'. 'Holoploid genome size' is a shortcut for the DNA content of the whole chromosome complement characteristic for the individual (and by generalization for the population, species, etc.) irrespective of the degree of generative polyploidy, aneuploidies, etc. This term was lacking in the terminology and is for reasons of linguistic consistency indispensable. The abbreviated terms for monoploid and holoploid genome size are, respectively, Cx-value and C-value. Quantitative data on genome size should always indicate the C-level by a numerical prefix, such as 1C, 1Cx, 2C, etc. The proposed conventions cover general fundamental aspects relating to genome size in plants and animals, but do not treat in detail cytogenetic particularities (e.g. haploids, hybrids, etc.) which will need minor extensions of the present scheme in a future paper.

Three independent databases of eukaryotic genome size information have been launched or re-released in updated form since 2005: the Plant DNA C-values Database (), the Animal Genome Size Database () and the Fungal Genome Size Database (). In total, these databases provide freely accessible genome size data for >10 000 species of eukaryotes assembled from more than 50 years' worth of literature. Such data are of significant importance to the genomics and broader scientific community as fundamental features of genome structure, for genomics-based comparative biodiversity studies, and as direct estimators of the cost of complete sequencing programs.

No abstract