■ INTRODUCTION TO THE METHOD 



extend its geographical range throughout Great Britain and over a part at least of the 

 continent of Europe (and perhaps farther afield) it is possible that it may be on the 

 way towards the production of yet other new forms; for the triploid hybrid between 

 Nasturtium uniseriatum and JV. officinale, which is also widespread in Europe, is quite 

 capable of giving rise to a variety of new types by strictly cytological means. In any 

 case the triploid itself must be recognized as a type of watercress which is necessarily 

 still younger than its parent species and which is at present spreading vegetatively. 



o J^ 



"o 



•» * ^ ^ ;y "^' 



^Srf^ 



4 



4 



^^-h 



4 



c 



Fig. 4. Explanatory diagrams to Fig. 3. x 2000. After Howard and Manton (1946). a. Diploid 

 watercress (Fig. 3c). b. Triploid hybrid (Fig. -^d). c. Wild tetraploid [Nasturtium iiniseriatiim 

 How. & Mant.) (Fig. 3^). d. Autotetraploid watercress (Fig. 3/). 



Looking again to the past, however, it may be suggested that the low-numbered 

 N. officinale itself may not be quite as simple as it seems, for a 'haploid' as high as n= 16 

 is unusual among its nearer relatives, the majority of which (in the genus Cardamine 

 and others) possess « = 8. It may therefore be suspected that Nasturtium officinale may 

 in turn have originated at some still earlier period and at a lower cytological level by a 

 repetition of the processes which have since produced N. uniseriatum. The last is specula- 



jV. uniseriatum may therefore have to be superseded on grounds of priority by one of two other possible 

 synonyms. For the sake of simplicity these emendations have for the moment been disregarded, since 

 they are at present not based on the same type of cytological evidence as that quoted in the text, 

 which, for convenience, follows Howard and Manton (1946), though without prejudice to the ultimate 

 adoption of Shaw's proposals should they be further authenticated. 



II 



