226 



TAXONOMIC STUDIES IN THE GENUS CERASTIUM : I 



F Group A. C. atrovirens = + 2336 



F Group C. C. semidecandrum = — 194 



Specimens with values of F above + 1071 should be placed in C. atrovirens and those 

 with values of F below + 1071 in C. semidecandrum. 



The results of these analyses can, for convenience, be made into a " key " as follows: 



1. Specimens with values of F above — 5270 (Groups A and B) 2. 

 Specimens with values of F below — 5270 (Groups B and C) 3. 

 Where = - 42, A2 = - 20, A3 = - 7, A4 = - 90, A5 = + 29 



2. Groups A and B 



Specimens with values of F above + 145 Group A. C. atrovirens. 

 Specimens with values of F below + 145 Group B. C. pumilum. 

 Where Aj = + 4, A2 = - 17, A3 = + 11, A4 = + 5, A5 = + 16. 



3. Groups B and C. 



Specimens with values of F above + 9177 Group B. C. pumilum. 

 Specimens with values of F below +9177 Group C C. semidecandrum. 

 Where A^ = + 88, X2 = + 97, A3 = + 291, A4 = + 62, A5 = + 58. 



To use this key is comparatively simple. The values of the characters, seed size, pollen 

 size, etc. must first be determined. These are converted to the coded values given in 

 the tables of critical values shown in Tables 1 to 5. 



These coded values are multiplied by the appropriate factor; the value of seed size 

 by and so on. Note that the appropriate value of A must be used for the different 

 analyses, these values being shown under the appropriate portions of the key. The value 

 of F obtained indicates which of the two more sensitive analyses should now be used. 

 No further measurements are necessary and the same coded values are used but the 

 appropriate values of A are used to obtain the value of F for that specimen. 



However, it must be borne in mind that any such method as that adopted here has 

 very definite limitations. Briefly, these are : (1) only the characters used in the analysis 

 are used in the determination and this must always be a poor substitute for the "eye" of 

 the born taxonomist who, although often using a larger number of characters subcon- 

 sciously in his determination, is seldom able to describe these characters adequately. 

 (2) That a number of measurements must be made before the key can be used. This 

 precludes its use in the field. However, where field characters can be used, the need 

 for such analyses does not arise. 



Nevertheless, the method has certain very definite advantages which may more than 

 outweigh the disadvantages mentioned above. For instance, it is possible to calculate 

 the probability of correct classification by the use of appropriate statistical tables. The 

 extent that a specimen approximates to another can also be calculated and this may be 

 of great use in separating back-crosses from parents or 1st generation hybrids, etc. To 

 those not gifted with taxonomic " sense," the method furnishes a means of correctly 

 determining specimens which otherwise would have to be sent to a specialist. Finally, 

 although the method depends in the first place on the subjective decision regarding the 

 choice of groups and samples to represent those groups, these subjective decisions are 

 subsequently tested objectively and the allocation of further unclassified specimens is 

 also objective. 



By means of this treatment of the data it is possible to obtain a clear separation 

 of these three taxa. Taking into account the differences which exist in distribution, 

 ecology, etc., it appears that the recognition of these three taxa as species is fully 

 justified. 



