214 



TAXONOMIC STUDIES IN THE GENUS CERASTIUM : I 



The cleft was measured from the deepest point of the cleft to the extreme point of the 

 longest of the petal divisions. 



6. Petal Length/ Breadth ratio. This ratio was that of the total length of the petal 

 to its breadth. Length was measured as described above and breadth was measured at 

 widest point at right angles to a line joining the lowest point 'of the cleft to the centre 

 of the petal base. 



The Character Seed Size 



This character was examined in detail because Scandinavian taxonomists have 

 attached much importance to it. A good example of its use is to be found in Lindman 

 (1918). It is used, in this flora, to distinguish between C. atrovirens and C. subtetrandrum 

 and between C. pumilum and C. glutinosum; the seed sizes being given as 0-7 mm and 

 0-5 mm. respectively. No seed size is given for C. semidecandrum. It is obviously necessary 

 to know to what extent this character varies from plant to plant in the same population 

 and from cloned plant to cloned plant under different conditions. 



The variation within the populations studied is shown in the histogram (Fig. 1). The 

 variation in response to growth conditions and between cloned plants under similar 

 conditions was investigated by the following experiment. 



Four plants grown from seed coming from one wild population were grown under 

 similar conditions. The seeds were collected and measured and the data are shown in 

 the histogram (Fig. 2) as Group 1. This group can be regarded as indicating the variation 

 between individuals in a single population. Eight rooted cuttings of a single plant from 

 the same population were treated as follows : four of these plants were grown in sand, 

 watered with tap water, and kept as dry as was possible without damaging the plants. 

 The variation in their seed size is shown as Group 2 in Fig, 2. This group may be 

 considered as showing the variation in seed size of a single genotype under one set of 

 extreme conditions. The other four were grown in a 50 per cent John Innes potting 

 compost - 50 per cent sand mixture - with an ample water supply. Group 3 of Fig. 2 

 shows the variation in the size of seeds produced by these plants and represents the 

 variation in seed size of the same genotype as Group 2 but grown under very different 

 conditions. From histogram 2 it can be seen that, although there is quite a wide range 

 of variation for each group, the range and distribution of this character is little affected 

 by differences of conditions of growth. It follows that this character is also fairly constant 

 in the population used, and a fairly small number of seeds can be used to sample the 

 variation in a wild population. 



It was found that the mean seed size differed for the three primary groups consisting 

 of specimens of C. atrovirens (Group A), C. pumilum (Group B) and C. semidecandrum 

 (Group C) as can be seen from the histogram (Fig. 1). It can also be seen that there 

 is considerable overlap of the variation of this character in the three primary groups, 

 there being a large area in the centre of the histogram where all three groups overlap. 

 Nevertheless, it appeared that the means of the primary groups might be significantly 

 different. Since N is large for each group the significance of the difference of the means 

 was determined by calculating the standard error of the difference between the means. 

 If the means are i A and z B, their standard deviations and and the numbers in the 

 samples and N2 respectively, then the standard error of the difference between the 

 means is given by the formula 



(1) 



