266 



P. F. YEO 



is offered to account for the achievement of fertiUty, in the diploid state, of forms whose 

 ancestry includes both diploid and tetraploid forms. Such a result has not been achieved 

 experimentally and it is therefore of interest to enquire whether a comparable process is 

 known in any other group. 



Anderson & Hubricht (1938) demonstrated, by morphological investigation, intro- 

 gression from the tetraploid Tradescantia canaliculata into the diploid T. hracteata. This 

 state of affairs is comparable with that described in some Euphrasia populations. Unfor- 

 tunately, the chromosome numbers of the forms in the population investigated were not 

 stated. The authors add that the case needs further investigation, since T. hracteata 

 presents the most complicated pattern of intraspecific variation yet encountered among 

 American Tradescantias. 



The second example is also provided by wild plants and is more fully investigated 

 than the first. It concerns hybrids between Polystichum lonchitis and P. lohatum, des- 

 cribed by Manton (1950). These two species are diploid and tetraploid respectively. 

 Plants of the putative hybrid, P. illyricum, from a locality where the other two species 

 grow together, formed an almost continuous series of morphological intermediates between 

 P. lonchitis and P. lohatum. Cytological investigation showed that some plants of 

 P. illyricum were triploid, but one, which resembled P. lohatum more closely, was tetra- 

 ploid, and one, which resembled P. lonchitis more closely, was diploid or nearly so. The 

 behaviour of the chromosomes at meiosis in the triploid showed that P. lohatum is an allo- 

 tetraploid, one of whose parents is P. lonchitis. The diploid plants, referred to P. illyricum, 

 thus appear to have originated in the way suggested for the Euphrasia hybrids discussed 

 in this paper. Manton says : " That some signs of hybridity still persist in such progeny 

 would merely seem to imply that some extra chromosomes belonging to one or other 

 species are still present or that a measure of gene exchange can occur between the chromo- 

 somes of the two species." It is not clear whether the second alternative concerns the 

 sets which pair at meiosis in the triploid or those that usually do not. However, one can 

 assume that the sets which pair in the triploid, while being cytologically homologous, are 

 not genetically identical. In that case, since there will have been crossing-over between 

 the chromosomes that paired in the triploid, characters found in the tetraploid will be 

 expected to appear in the back-crossed diploid. 



Miintzing (1937) obtained a partially fertile diploid in Dactylis by crossing the 

 tetraploid D. glomerata with the diploid D. polygama (D. aschersoniana) and back-crossing 

 the resulting triploid with the diploid parent. In this case D. glomerata itself forms 

 quadrivalents frequently at meiosis and the triploid forms a high proportion of trivalents. 

 Pollen fertility in the triploids was between 4 and 10 per cent. This indicates a closer 

 relationship between the haploid set in the diploid with both sets in the tetraploid than 

 is the case in Euphrasia. 



The cross Viola reichenhachiana X riviniana provides a final example. Valentine 

 (1950) crossed these two species, the first of which is a tetraploid with 2n = 40 and the 

 second a diploid with 2n = 20, and thereby produced a triploid. 



The commonest arrangement of the chromosomes at meiosis consisted of ten bivalents 

 and ten univalents. The triploid was highly sterile, but most of its few offspring had a 

 diploid chromosome number between 18 and 21. These varied from weak and non- 

 flowering, through flowering but sterile, to slightly fertile. They were produced cleisto- 

 gamously, so that the process involved differs from that suggested for Euphrasia 

 (p. 255), but the selfing of the triploid is an alternative possibility for Euphrasia that has 

 been mentioned (p. 256). The fertility of the diploid is very low, but might increase 

 in later generations. It is interesting to note that Valentine, in his discussion, considers 

 the possible achievement of fertility in hybrids between species differing in chromosome 



