HYBRIDISATION OF EUPHRASIA 



259 



The three forms concerned were differentiated ecologically, as well as morpho- 

 logically. Lists of species occurring with E. anglica and E. vigursii at Roborough are 

 given in Table 3. E, anglica occupied its characteristic habitat of short turf. E. " nemo- 

 rosa " was usually found in longer grass than E. anglica, round bushes of Ruhus fruticosus 

 agg. and Ulex europaeus, but it was also present in quite short grass by the roadside in 

 places. E. vigursii, also a tall plant, grew in still taller vegetation. This wa^ characteristic- 

 ally a heathy community consisting of Agrostis setacea, Calluna vulgaris. Erica cinerea, 

 and Ulex gallii. It also grew in a grassy community with rather long grass, mostly Agrostis 

 setacea, and more or less scattered fronds of Pteridium aquilinum. The first of these two 

 communities occupied the greater part of the heath and E, vigursii is probably commoner 

 than either of the other species. 



The habitat differences keep the greater part of the populations of the forms apart. 

 Occasionally, however, E. anglica and E. " nemorosa " were more or less intermixed. At 

 such places no intermediate could be found. E. vigursii mixes less with these two than 

 they do with one another, but in some places it niet E. anglica, and, in intermediate habitats 

 on the embankment of a side-road, plants intermediate between the two were found. 



Interpretation of the Situation at Roborough Down 



E. vigursii occurs at Roborough in company with diploid and tetraploid forms between 

 which it might, on morphological grounds, be interpreted as a hybrid. It is a fertile 

 diploid, differentiated in habitat from its presumed diploid parent. This situation appears 

 to be analogous to that at Withypool, and the same hypothesis can be used to explain it. 



There are, however, some differences between the Withypool and Roborough 

 situations. At Roborough, E. vigursii is not so clearly intermediate between the two 

 supposed parents as E.187A is. This can be explained by the complex series of crosses 

 involved, causing the removal of genes from the tetraploid into a new genetic environment 

 in which they may have a modified action. 



E. vigursii is morphologically distinct from the fertile hybrid at Withypool. But, 

 at least at localities other than Roborough, it appears to have E. micrantha as its tetraploid 

 parent. The morphological difference between the two forms could be attributed to the 

 transmission of different sets of genes from the tetraploid. Another possibility is that, 

 as already suggested, E. confusa has been involved in giving rise to the hybrid at Withypool, 

 as well as E. micrantha. 



Though the tetraploid accompanying E. vigursii at Roborough is not true E. micrantha, 

 the latter might nevertheless have been its original parent there, since E. vigursii, being 

 fertile, can persist in the absence of its parents. E. micrantha may have subsequently 

 died out on the heath, or it may still occur somewhere on the large part of the heath not 

 explored. 



The habitat which E. vigursii occupies at Roborough is different from that of E.187A, 

 and is more clearly distinct from the habitat of E. anglica. In addition, E. vigursii forms 

 an abundant population spread over a wide area. The importance of the ecological 

 conditions in enabling the diploid population to become split into two different forms 

 is again evident. 



Further Examples of E. vigursii Populations 



Some other populations of E. vigursii at more westerly localities will now be described. 

 Species lists for the first four dealt with are given in Table 3. The habitats were similar 

 to that of E. vigursii at Roborough but poorer in species, probably because of burning. 



