THE CYTOLOGY OF BRITISH SPECIES OF EUPHRASIA 



107 



another. They possess rounded, cuneate-based leaves with rounded teeth, and the 

 former includes both eglandular and long-glandular forms. The mere discovery of 

 other diploid groups would not, of course, identify that which was concerned in giving 

 rise to the European tetraploids. 



Endemism 



Euphrasia shows a high rate of endemism. Thus of the species counted, E. foulaensis, 

 E. marshallii, E. occidentalis, E. confusa, E. pseudokerneri, E. rivularis, and E. anglica are 

 endemic to the British Isles, or almost so. {E. occidentalis occurs in Brittany and probably 

 in Belgium, and E. confusa and E. foulaensis occur in the Faeroes.) The cytological 

 work shows that in these cases variation in chromosome number is not the cause of the 

 multiplicity of species, many of which are very localized in distribution. Uniformity 

 of chromosome number appears to be the rule, so that it seems improbable that any 

 other British endemics owe their existence to the possession of different chromosome 

 numbers. 



Summary 



Chromosome counts at pollen mother cell meiosis were made in fifteen British 

 Euphrasia forms, covering thirteen species. 



The method used is outlined and the selection of material for fixing described. 



The numbers n=22 and n = ll were observed. All previously reported numbers 

 are listed, and a table of all known numbers is presented (p. 105). In addition to the 

 regular numbers, E. hrevipila var. notata showed pairing failure in one pair of chromosomes, 

 and plants were found with 2n=444-l, presumably caused by this pairing failure; also 

 a single triploid individual was observed, which most probably forms 11 bivalents and 

 11 univalents at metaphase. This plant was a hybrid between E. micrantha and E. anglica. 



The relation of chromosome numbers to classification is discussed. The division 

 between diploids and tetraploids coincides with one of the divisions of Pugsley's classi- 

 fication, the Series Hirtellae being diploid and other groups tetraploid. The raising of 

 the Hirtellae to the rank of Subsection is recommended on morphological as well as 

 cytological grounds. The anomalous forms E. hrevipila var. notata and var. reayensis are 

 tetraploids; this confirms their exclusion by Pugsley from the Series Hirtellae. 



The existence of diploid and tetraploid series makes it possible for certain pairs 

 of species to grow in company and remain distinct. This they frequently do; triploids, 

 however, are very rare. 



The cytology of the triploid shows that E. micrantha is an allotetraploid. In con- 

 sidering the origin of tetraploids, groups should be probably considered as a whole, since 

 the species they comprise are closely related to one another. E. anglica is perhaps not 

 a direct ancestor of a tetraploid, and E. micrantha is probably derived by divergence 

 from a primitive tetraploid. 



A chromosome survey of other groups of Euphrasia, particularly those within the 

 Section Semicalcaratae, is needed. 



Differences of chromosome number are not the cause of the multiplicity of Euphrasia 

 species. 



Acknowledgments 



I am grateful to Professor T. G. Tutin for his interest and advice during the course 

 of this work, and for criticizing the manuscript of the paper; also to the Research Board 

 of The University College of Leicester for the provision of a Research Scholarship, and 

 grants for travelling expenses. I have also been helped by Mr. Martin Borrill with 



