TAXONOMIC STUDIES IN THE GENUS CERASTUIM 



325 



been calculated in order to distinguish between C. atrovirens and C. suhtetrandrum and the 

 results are given below. The specimens of C. suhtetrandrum used in this analysis were all 

 from Scandinavian localities and many of them were originally determined by Murbeck 

 himself. The C. atrovirens were determined by the preceding discriminant analysis. The 

 characters used were :- 



(1) Seed size as defined before (x) Table 2 (Whitehead 1955) 



(2) Pollen size as defined before (3;) Table 1 (Whitehead 1955) 



(3) Length/Breadth ratio of petals (h) (see Table 1). 



Table 1 



Critical values of the ratio petal-length/petal-breadth 



01 



2-518 



21 



2-878 



41 



3-238 



61 



3-598 



81 



3-958 



02 



2-536 



22 



2-896 



42 



3-256 



62 



3-616 



82 



3-976 



03 



2-554 



23 



2-914 



43 



3-274 



63 



3-634 



83 



3-994 



04 



2-572 



24 



2-932 



44 



3-292 



64 



3-652 



84 



4-012 



05 



2-590 



25 



2-950 



45 



3-310 



65 



3-670 



85 



4-030 



06 



2-608 



26 



2-968 



46 



3-328 



66 



3-688 



86 



4-048 



07 



2-626 



27 



2-986 



47 



3-346 



67 



3-706 



87 



4-066 



08 



2-644 



28 



3-004 



48 



3-364 



68 



3-724 



88 



4-084 



09 



2-662 



29 



3-022 



49 



3-382 



69 



3-742 



89 



4-102 



10 



2-680 



30 



3-040 



50 



3-400 



70 



3-760 



90 



4-120 



11 



2-698 



31 



3-058 



51 



3-418 



71 



3-778 



91 



4-138 



12 



2-716 



32 



3-076 



52 



3-436 



72 



3-796 



92 



4-156 



13 



2-734 



33 



3-094 



53 



3-454 



73 



3-814 



93 



4-174 



14 



2-752 



34 



3-112 



54 



3-472 



74 



3-832 



94 



4-192 



15 



2-770 



35 



3-130 



55 



3-490 



75 



3-850 



95 



4-210 



16 



2-788 



36 



3-148 



56 



3-508 



76 



3-868 



96 



4-228 



17 



2-806 



37 



3-166 



57 



3-526 



77 



3-886 



97 



4-246 



18 



2-824 



38 



3-184 



58 



3-544 



78 



3-904 



98 



4-264 



19 



2-842 



39 



3-202 



59 



3-562 



79 



3-922 



99 



4-282 



20 



2-860 



40 



3-220 



60 



3-580 



80 



3-940 



100 



4-300 



From this analysis the values of A were found to be :- 



Aj = + 32-9449 

 ^2 = - 84-5845 

 A3 = + 25-8831 



The mean values of F were as follows :- 



F C. atrovirens = — 41-8732 

 F C. subtetrandrum= — 2280-2617 



It follows therefore that specimens for which a value of F greater than — 1,161 should 

 be allocated to C. atrovirens and those for which the F value is less than — 1,161 should 

 be allocated to C. suhtetrandrum. The distribution of the values of F for the specimens 

 used is shown in Fig. 1. It was found that complete separation was obtained and that 

 there is complete bimodality in the distribution of F values. 



An examination of the values of F for the specimens of C. atrovirens is of interest 

 since it is found that those with the lowest values, i.e. those specimens which in respect 

 of the characters used are most like C. suhtetrandrum, are from localities in the North 

 of Britain, where C. suhtetrandrum has been from time to time reported as occurring. An 

 examination, however, of British material determined as C. suhtetrandrum showed that 



