Peoblems of Hetekogeneity 189 



one, in which the values and behaviors of the individual qualities and attributes 

 are lost sight of or ignored. 



In heterogenetic phenomena, as elsewhere, the unit-character conception seems 

 unnecessary and a cumbersome method of reasoning, and it is essentially vitalistic 

 in its conception. In physical phenomena specific properties are present or 

 absent, are changed with sharp alternativeness, and no need is felt of either such 

 concepts or for inventing mythical pangenes or biophores. If such concepts are 

 not needed in physical science, why introduce them into biology ? 



For the purpose of this paper I shall consider first those attributes which, as 

 far as known, are not capable of being divided into lesser ones, or simplest char- 

 acters, and from these shall proceed to the consideration of characters of increas- 

 ing complexity. It must be understood that this method of attack does not 

 imply in any degree a preference for any current hypothesis. I shall begin with 

 the simplest characters that can be certainly and generally recognized. The 

 criterion for the recognition of this condition has already been stated, and while 

 the criterion will always remain the same, the condition it defines will change 

 from time to time. A name is needed for the conditions from which I start, 

 and throughout this paper I shall call them the " simplest characters," implying 

 no more than the idea already expressed, that they are the simplest conditions 

 that can now be certainly determined and profitably studied among the multi- 

 tude of specific properties, attributes, and conditions presented by organized 

 bodies. 



The development of the " pure-line " hypothesis of Johannsen introduces 

 further complications into the problem. A basal unit or " gene " comparable 

 with " unit-characters," " unit-factor," " element," " allelomorph," and so on, 

 and mainly an expression of the results of the experience of neo-Mendelian 

 hybridology, is conceived of as having a fixed value or manifestation which 

 behaves as a constant factor in organic evolution. The combination of genes in 

 the organic form into a harmoniously acting state of stability, a genotype, is 

 essentially of the same order as De Vries's conception of the constitution of 

 species by combined elementary species. The isolation of genotypes by Johann- 

 sen and others, of clones by Jennings, leaves no doubt that there are in 

 the population states of stability which can be separated and maintained as pure 

 homozygous-acting strains. These are always to be distinguished by pedigree 

 breeding methods and are different from " phenotypes," which are separated by 

 purely descriptive methods and may or may not represent " real things." 



The genotype conception, while it expresses many experiences in language that 

 is not liable to be misconstrued, nevertheless introduces complications. Opinion, 

 supported by practice in line-breeding, has rapidly established the view that 

 genotypical constitution changes by jumps alone, without intergrades, and that 

 the obvious and constantly present overlapping of genotypes when isolated is 

 only the " fluctuations " of the genodifferences present. The added current 

 belief that no selective accumulation of the fluctuations of the differentiating 

 genes can alter the mean of the " gene," results in much of the work in selection 

 being interpreted as due to the isolation, by " selection," of " genotypes." 



With respect to the " gene," as with the " unit-character " or the " simplest 

 character," the essential present need is more real knowledge of these characters. 

 It is necessary to determine whether their commoner differences are mere quanti- 

 tative oscillations about a mean, whether their rarer jumps to new states are 



