62 THE VARIATION OF ANIMALS IN NATURE 



(e.g. at the lower limit, the intervention of mutual infertility). 

 But, as organisms diverge in many characters, and as these 

 are not correlated in any universal scheme of divergence, any 

 attempt to fix a downward limit fails. 



The first objection is far more cognate to our problem. 

 The universal occurrence of individual variation has led 

 certain writers to assert that the individual is the only real 

 unit and that species and similar groups are devoid of any 

 significance. This view is worth dwelling on for a moment, 

 as its importance is not fully recognised. Finding agreement 

 between the members of his species in a limited number of 

 characters the systematist has perhaps given undue prominence 

 to them. When the term similarity is introduced into the 

 definitions of systematic units, we may well ask if any two indi- 

 viduals, even of a moderately complex phylum, are ever alike in 

 all their characters (cf. p. 60, supra) . If this is never the case, we 

 may also ask how it is that any discrete groups, such as species, 

 have come to be recognised and what may be the value of 

 a classification that recognises such crude groupings. The 

 answer to this may be given briefly. In spite of very extensive 

 individual variation (a great part of which is of unknown 

 hereditary status and may be non-heritable), the systematist 

 tends to find certain regular correlations, associations of a 

 limited number of characters that occur regularly in individuals, 

 and it is this correlation that, amid a very great amount of 

 individual variation, constitutes the basis of species-diagnosis. 

 Such correlations are, of course, of very varying intensity and 

 can involve a greater or less number of characters of various 

 kinds ; but, though they cannot be standardised as a univer- 

 sally recognisable grade, the taxonomic procedure is justified. 

 It is necessary to make the proviso that a number of species 

 in each group are founded on inadequate statistical data. 

 Indeed so great is the disparity between the number of species 

 described by the systematist and the knowledge of natural 

 variation of the populations from which species are abstracted, 

 that some systematists (e.g. Ramsbottom, 1926, p. 28) have 

 been impelled to draw a distinction between ' the natural 

 species ' and ' the taxonomic species,' and one of the authors 

 of the present volume has suggested that forms which, by 

 reason of the poverty of material, imperfect preservation, 

 or the lack of adult specimens, are of uncertain status, though 



