HYBRIDISATION OF EUPHRASIA 



265 



supposed hybrid was E. nemorosa, which was abundant. No pure E. anglica was found 

 at Friday Street, though it is common in Surrey. The apparent hybrid was found to 

 occupy only a few square yards of ground, over which it occurred quite densely in short 

 grass. It differed from E, anglica primarily in its small, narrow, ovate leaves. The corolla 

 lobes were unusually narrow for E. anglica ; this might have resulted from hybridisation 

 but it was not a character of the E. nemorosa population here. The colony was visited 

 in July and late September, and on neither occasion were the flowers large. The difference 

 between my specimens and Salmon's in this respect may be attributed either to the 

 evolution of a smaller flower-size in the last fifty-five years, or to my not having found 

 the same population as Salmon. 



Friday Street is on acid sandy soil, unsuitable for E. pseudokerneri, which is not 

 likely to occur less than two miles away, on the North Downs, so that it seems as if the 

 tetraploid parent must be E. nemorosa. 



E. anglica X confusa occurs quite frequently. These two species are both dwarf in 

 habit, so that hybridisation results in no change in stature. E. anglica affected by intro- 

 gression from E. confusa tends to show the flexuous branches of the latter, but the main 

 indication of hybridity is see in the leaf shape. The occurrences of this hybrid known 

 to me are : 



V.c. 5, S. Somerset : round Withypool, visited in 1952, E. anglica was almost 

 invariably accompanied by E. confusa, and in some populations it appeared to be modified 

 by introgression from that species. At Elworthy in 1951 Mr. M. C. F. Proctor collected, 

 as well as more or less normal E. anglica, plants that were strongly modified, especially 

 in leaf shape. E. confusa was collected at the same place. 



V.c. 6, N. Somerset : at Gorsey Bigbury, N.E. of Cheddar (grid ref. 31/479559), 

 E. anglica and E. confusa were found growing together, and the former showed definite 

 characters of the latter. Herbarium specimens collected by Turrill (in Herb. Kew) 

 from Mendip Moors, near Cheddar, 1930, appear to represent this hybrid, as do those 

 of Bucknall from " boggy ground, in thick spongy turf on Mendips, near Rowberrow," 

 as *' E. rechingeri, with the parents," 1916. The parents of E. X rechingeri Wettst. are 

 E. kerneri Wettst. and E. rostkoviana, and Bucknall (1917), citing earlier gatherings 

 from here, said it occurred with these two species. His description of " E. kerneri " 

 suggests that it was E. confusa, and E. anglica was usually called E. rostkoviana at that 

 period. Pugsley (1930, 540) regards the Rowberrow plant as abnormal E. anglica. 



V.c. 39, Stafford, and v.c. 57, Derby : at Stanshope, near Wetton (v.c. 39) (grid ref. 

 43/126544) and at Thorpe (v.c. 57) (grid ref. 43/151506), populations of E. anglica were 

 found which showed, in their habit and foliage, indications of introgression from E. confusa, 

 which was growing with it in both places. 



E. anglica is a very variable species, in spite of its nearly always being easily identified, 

 and part of this variation is probably due to introgression from various tetraploid species. 



In the British Isles, the diploid E. rostkoviana is usually accompanied by the tetraploid 

 E. hrevipila, and some of its variation may be due to introgression from that species. A 

 probable indication of this is the variation in capsule shape, which is typically short and 

 broad but is sometimes longer and narrower. The two species are similar in habit, both 

 being tall. 



Much Continental material of E. rostkoviana shows characters recalling those of 

 tetraploids, but field observation is needed to assess the significance of this. 



Discussion 



The main conclusion drawn from the observations described in this paper is that 

 genes can pass from tetraploid species of Euphrasia into diploid species. A hypothesis 



