64 'WHAT IS A SPECIES?' 



of introduction, setting forth the main conclusions which 

 will be defended. 



Diagnosis, it will be maintained, is founded upon the 

 conception that there is unbroken transition in the 

 characters of the component individuals of a species. 

 Underlying this idea are the more fundamental con- 

 ceptions of species as groups of individuals related by 

 Syngamy and Epigony. 



In immense numbers of cases it will be shown that 

 the component individuals of a species do not form 

 an unbroken series, but one that is sharply broken at one 

 or more points. At each of these breaks the older syste- 

 matist made a new species, which the modern systematist 

 has rejected, because in his day the more fundamental 

 criteria have been actually established or by strong 

 indirect evidence have been inferred. When the test 

 of Diagnosis necessarily fails — as it will be shown to 

 do in many large groups of examples — the appeal is made 

 to Syngamy and Epigony. 



Syngamy and Epigony are but two sides of the same 

 phenomenon — Reproduction. Although occasional union 

 between individuals of distinct species may occur in nature, 

 sometimes leading to the production of hybrid offspring, 

 this is not the ' free interbreeding under natural con- 

 ditions ' which I have called Syngamy. Syngamy, thus de- 

 fined, implies the production of normal offspring capable of 

 continuing the species — implies Epigony. As a practical 

 criterion, the evidences of Syngamy are generally much 

 easier to collect than those of Epigony. Both Syngamy 

 and Epigony can be established by indirect evidence 

 based on a sufficient number of accurate observations 

 upon the habits and modes of occurrence of individuals. 

 The criterion of Syngamy of course fails in the case 

 of parthenogenetic and self- fertilizing species. In such 

 cases, like that of the Bee Orchis referred to below on 

 p. 92, we are compelled to fall back on Epigony. This 

 latter criterion may lead, although only in rare and 

 exceptional instances, to an erroneous inference, when 

 hybrid are mistaken for normal offspring. 



There is a great advantage in the admission that 



