VARIATION. ADAPTATION. ORIGIN OF SPECIES 387 



three (DE VRIES, 1901 a) (compare also KLEBS, 1903, as to the types of varia- 

 tions and as to the origin of species), fluctuating variations, adaptive variations, 

 and mutations). To these may be added a fourth type, variations which arise 

 from hybridization ; into these, however, we cannot enter here. 



Fluctuating variations may be termed also individual, because they show 

 themselves in single members of the species, arising apart altogether from 

 crossing or extraneous influences of that sort. If the seeds from a single 

 capitulum of a member of the Compositae be planted in a garden the plants 

 which result vary extremely in weight and size, and, later, the different organs 

 of each plant also vary in number, size, and weight from the quantitative aspect. 

 If statistics be collected of a larger number of individuals a certain average 

 may be determined for each character, and the deviations from that mean occur 

 less and less frequently the greater they are. We may, in fact, from the data 

 so established construct a curve (Galtonian curve) which corresponds more or 

 less exactly to the law of probabilities (QUETELET). The curve has one crest 

 decreasing rapidly to zero on either side. These individual variations may be 

 illustrated first by a few examples taken from observations on wild plants 

 selected at random. 



Number of rays in the archegoniophore of Marchantia (LuDWlG, 1900, 

 p. 22) : 



No. 78 9 10 ii 12 13 



Frequency i 14 307 152 44 3 i Total 522. 



Number of petals in Linaria sfiuria (VOcHTiNG, 1898) : 



No. 23456789 



Frequency in actinomorphic flowers i 2 43 810 52 a i i Total 912. 

 > zygomorphic ,, 4 240 60250 169 7 I Total 60671. 



Length of seeds of Phaseolus vulgaris (DE VRIES, 1901 a, p. 34): 



Mm. 8 9 10 ii 12 13 14 15 16 

 Frequency i 2 23 108 167 106 33 7 i 



Percentage of sugar in 40,000 beets from Naarden (DE VRIES, 1901 a, p. 74): 



% sugar 12 i2| 13 13! 14 141 15 15^ 

 Frequency 340 635 1192 2205 3597 5561 7178 7829 



% sugar 16 i6| 17 17! 18 i8 19 

 Frequency 6925 4458 2233 692 133 14 5 



The data given in this last example are also expressed in the form of a 

 curve in Fig. 114. It should be noted that curves with several maxima occur, 

 and also half curves as well, but these cannot be further discussed. 



We have now to discuss what is the relationship of these individual varia- 

 tions to species formation. First of all, we must note that there is a very close 

 connexion between such variations and the mode of formation of agricultural 

 breeds or races by selection. Thus, in the sugar-beet industry, the selection 

 and employment for propagation of the seeds of those plants which are richest 

 in sugar leads to a marked rise in the general average of the sugar percentage in 

 the beet. Fifty years ago the percentage was 7-8 per cent., now it has been 

 raised to about 15 per cent. In the same way, by rigorous selection, races 

 may be produced which will exhibit especially large flowers or fruits, better 

 flavour, increased succulence, &c. So far as we know, no new characters arise 

 spontaneously, although those already existent may be added to or reduced. 

 The limits of such variations are usually reached in a few generations (3-5) ; 

 further selection merely serves to fix the character which has been acquired. It 

 must be specially noted, however, that such characters are liable to fade away 

 quite as rapidly as they appear ; after a few generations, if selection be dis- 

 continued, the original condition is reverted to. Herein lies a great distinction 



c c 2 



