NUPHAR INTERMEDIA IN BRITAIN 



21 



given in Table 1. It is also notable that in N. lutea the range quoted by Caspary for 

 continental plants (13-26) differs in the same direction from the British range of 11-20 

 in the aggregate sample (83 plants). 



Table 6 



Meristic data for floral parts from Continental European Nuphar taxa (data after Caspary 1870). 



Taxon 



Petal number 



Stamen number 



Carpel number 



Total number 

 floral parts 



N. pumila 



10-71 ± 0-16 



46-38 ± 1-03 



10-03 ± 0-05 



69-08 ± 1-42 



N. intermedia 



18-16 ± 0-85 



94-51 ± 2-20 



13-18 ± 0-31 



130-60 ± 2-87 



N. lutea 











(a) midpoint 



19-50 



146-50 



17-00 



190-00 



(b) range 



13-26 



86-207 



10-24 



130-250 



In addition to his data on meristic variation in Nuphar, Caspary (1869, 1879) provided 

 much information on seed and pollen fertility. For plants referred to N. intermedia he 

 reported a range of pollen fertilities from 14 to 71 per cent, these being in contrast with 

 N. lutea and N. pumila, in neither of which did he obtain plants with pollen fertility 

 lower than 95 per cent. Seed set was correspondingly lower in the N. intermedia examined 

 by Caspary, but was surprisingly variable, the mean number in colonies observed by him 

 ranging from 5-6 to 41-7. This compares with a range of 32-7-135-1 mean numbers per 

 fruit for N. pumila colonies and 93-8-418-1 for N. lutea colonies. 



One of Caspary's (1869, 1870) most important contributions to the understanding 

 of N. intermedia was to cross N. lutea and N. pumila reciprocally and bring the hybrids 

 to flower in cultivation. Morphological comparisons of these hybrids with wild N. 

 intermedia convinced him that the latter had originated in the same manner in nature. 

 Pollen counts in the artificial hybrids showed 86-3 per cent malformed pollen (N. lutea 

 female parent) and 85-6 per cent (N. pumila female parent). The corresponding average 

 numbers of seeds per fruit were 14-9 and 18-4. 



Discussion 



The status of the presumed parents 



The evidence presented above suggests that while both N. lutea and N. pumila may 

 show considerable local variability, they form reasonably homogeneous units throughout 

 their European areas and do not intergrade to any appreciable extent. The geographical 

 distributions of the two are at present not completely mutually exclusive, but, generally 

 speaking, N. pumila has a more northerly range than N. lutea, and also tends to replace 

 N. lutea in mountainous regions (Meusel, 1943). However, geographical and ecological 

 isolation cannot be looked upon as the only, or even the main, factors holding the two 

 species apart today, since, as we have seen, the experiments of Caspary, in which he 

 crossed reciprocally typical representatives of the two, showed that, while they are inter- 

 fertile enough to produce a vigorous Fi, the Fj itself is of relatively low fertility. There 

 is thus a partial inherent barrier to gene exchange in spite of their possession of the same 

 chromosome number, and N. lutea and N. pumila must be looked upon as ecospecies 

 in Turesson's terminology. It remains probable, however, that the original differentiation 

 of the two species from their common ancestor took place under conditions of spatial 

 isolation, and that the present distributional trends - N. pumila towards the north and 

 N. lutea more southerly - in Europe reflect former complete vicariousness. There is no 



