examination. Without any doubt it is necessary in such a case 

 first to establish — if possible — once and for all by means of 

 careful studies in heredity whether this or that form is constant. 

 By this I do not mean, whether the form in question always 

 occurs in the type -locality with these or those distinguishing 

 features, but whether the qualities which characterise the form 

 from all others persist under other external conditions. 



Investigators have, as the reader is aware, also tried to 

 establish by experiment what factors can influence the formation 

 of forms and to what degree this influence takes place. But the 

 results arrived at will generally be, for various reasons, more or 

 less misleading. On this point I share Dahl's opinion (Grund- 

 lagen einer okologischen Tiergeographie, Jena 1921, p. 3) when 

 he says: 



"Noch unsicherer wird es, wenn man mit Tieren in der 

 Gefangenschaft experimentiert. um die Wirkung der einzelnen 

 okologischen Faktoren, die fur das Vorkommen in der Natur 

 mafsgebend sein konnen, festzustelleu, wenn man z. B., um fest- 

 zustellen, welche Temperaturschwankungen eine Tierart vertragt, 

 diese in der Gefangenschaft verschiedenen Temperaturen aussetzt. 

 Schon die Gefangenschaft an sich iindert namlich die Lebens- 

 bedingungen von Grund aus, so dafs man immer im Unklaren 

 bleibeu wird, wie weit das sich ergebende Resultat auf die ver- 

 anderte Temperatur zuriickzufiihren ist, wie weit das Experiment 

 also fiir das Vorkommen in der Natur Giiltigkeit besitzt." 



Banner man (Bull. Brit. Orn. Club, vol. 35, 1915, p, 136) 

 quotes a letter from Prof. Punnet, in which this great exponent 

 of Mendelism writes: "This problem of local races and adaptation 

 is one of the nuts left for us to crack — if we can — . The first 

 thing to be settled is whether the variation is genetic or whether 

 it is merely a direct response to a change in the environment 

 and only endures as long as the change in the environment 

 endures." 



Before this is experimentally established the majority of 

 speculations and conjectures about the affinity and constancy of 

 the forms can scarcely be looked upon as anything else but 

 "beating the air". 



The question of heredity and variation is not solved by 

 means of morphological comparisons, however accurate they may 

 be. Thus I share Stresemann's opinion (p. 293) that as long 

 as it is impossible to represent in figures the relation of the sub- 

 tile forms to closely allied forms, it is better to employ the old 



signs: > and < and to use ^ instead of = for a bird which 



evidently stands between two others. 



As regards the systematical order I have followed Reiche- 

 now's system in "Vogel Afrikas". Many writers have other 



