CYTOLOGICAL AND HYBRIDIZATION STUDIES IN THE GENUS DIANTHUS*

Abstract

The genus Dianthus has for long defied attempts to produce a wolkable classification. Up to the present only Williams (I893) has given a complete treatment of the genus. The characters by which he groups the species into sections and sub-sections are frequently inconstant and unreliable; in places he ignores his sectional characters with startling results. Thus D. colensoi Will. and D. mecistocalyx Will. were described from the same specimen yet placed in separate sections. For a fuller criticism Vierhapper (I898) may be consulted. Williams's treatment is unsatisfactory and needs to be revised. Several excellent monographs of small sections within the genus have appeared, notably Vierhapper (I898) and Novak (I927), and what is needed now is a framework into which the various species can be fitted. Rohweder (I934) has attempted this, using as a basis for delimiting sections ('groups') the ratio Chromatin Volume/Nuclear Volume. He substantiated his results with other quantitative measurements but, unfortunately, these were conducted on statistically unsatisfactory material. He applied his results too rigidly by arranging the species in a descending scale of chromatin/nuclear volume ratio. This method separated obviously related species such as D. liburnicus Bartl. and D. carthusianorum L. He also gave results to show that hybrids have ratios intermediate between those of the parents. Thus, if a species is a natural hybrid it is possible, by using Rohweder's method, for it to be assigned to a totally alien 'group'. There is little doubt that Rohweder's results are useful but they must be correlated with other investigations. One of the reasons for the genus Dianthus being so difficult to analyse taxonomically is the presence of hybrids which blur systematic boundaries. It is generally illogical and impractical to place taxa which produce fertile hybrids in separate intrageneric, supraspecific categories as Williams has done for D. collinus Wald. & Kit., and D. seguieri Vill. The systematic status of hybrids is only satisfactorily solved when sterility barriers are considered whilst delimiting sections, i.e. when the comparium or coenospecies is used as a basis for these categories. It is important to remember in this connection that new 'species' may arise by 'introgressive hybridization', whilst remaining diploid. In addition there are the incompletely investigated instances of 'Lotsyism' and the production of fertile 'species' from infertile hybrids by chromosome doubling. Another reason for using the comparium to delimit sections is that related species have a relatedgenetic composition and, therefore, a better chance of producing hybrids than those with dissimilar genetic complements. Incompatibility mechanisms of the type found in Prunus, etc., would upset this reasoning to a greater or less degree; so far they have not been found in Dianthus.