20 



YOLANDE HESLOP-HARRISON 



not significantly different, although the difference in the diameter of the stigmatic disc 

 is certainly so. 



Sepal length - 



« 1 1 1 1 — r— r " — r 



12 lb 20 24 Za 31 



Sepal width^mm^ 



Fig. 12. The relation of sepal length to sepal width in European Nuphar taxa. Code as in Fig. 11. 



Table 5 



Meristic and size data for floral characters in Continental European Nuphar taxa {ex herhariis; localities as 



in Table 3). 



Taxon 



Sepal 



Petal 



Carpel 



Peduncle 



length 



width 



length 



number 



diameter 



diameter 



N. pumila 

 N. intermedia 

 N. lutea 



17-93 ± 0-31 

 19-63 ± 0-37 

 25-91 ± 0-63 



10-02 ± 0-21 

 13-74 ± 0-27 

 21-50 ± 0-50 



5-85 ± 0-39 

 7-27 ± 0-22 

 14-30 ± 0-39 



9-85 ± 0-22 

 11-96 ± 0-20 

 14-93 ± 0-20 



5- 43 ± 0-21 



6- 07 i 0-13 



7- 85 ± 0-17 



3-48 ± 0-08 

 3-84 + 0-09 

 6-05 ± 0-23 



{All size measurements in mm.). 



Robert Caspary, who conducted extensive researches on the Nymphaeaceae in the 

 period 1855 to 1891 , accumulated a very large amount of statistical data on both Nuphar 

 and Nymphaea in Europe, and must be considered to be the first botanist to recognise 

 the importance of population studies in investigating the variability of these genera. The 

 counts of petal number, stamen number and carpel number which he published (1870) 

 for plants of N. pumila and N. intermedia have been condensed in the data given in 

 Table 6. The " N. pumila aggregate " in Table 6 refers to collections of the species 

 from six stations in the Vosges and Black Forest. The N. intermedia sample was also 

 derived from localities in these areas. For N. lutea Caspary reports only observed ranges 

 and these are given (together with their mid-points) in Table 6 for comparison. Caspary's 

 N. pumila aggregate differs little in carpel and petal number from the British N. pumila 

 aggregate, for which data are given in Table 1. While his data suggest a reasonable corres- 

 pondence in carpel number between the N. intermedia samples he examined and that 

 of the Chartners Lough sample, there is clearly a considerable difference in petal number, 

 the mean in Caspary's sample being almost twice that of the Chartners sample, and 

 actually greater than that for the aggregate British N. lutea sample, for which data are 



