Nature of Hereditu and Problem of Mechanism 315 



the hereditary attributes of the organism are causally ex- 

 plained by tlie chromosomes. 



Of tJie many definitions the one that most nearly expresses 

 the conception that will pervade this discussion is tliat given 

 by W. E. Castle: "By heredity, then, we mean organic re- 

 semblance based on descent." " The commendable things 

 about this definition are its non-commitment to any theory, 

 and the fact that it puts resemblance in tlie front line along 

 with the recognition that resemblance is due to descent. 

 Any adequate definition of heredity must hold the phenom- 

 enon of resemblance always in clear sight, and this in spite 

 of the fact that in the Mendelian mode of inheritance the 

 resemblance may skip one or mo.re generations. 



Our further discussion will fall under two heads. First, 

 we shall make a wide survey of resemblance due to descent 

 for the purpose of learning how far its connection with 

 chromosomes actually extends ; whether in a word, the con- 

 nection is a universal principle. Second, we shall then have 

 to see what we are justified in supposing to be the nature of 

 the connection. 



It will be noticed that the first statement admits in ad- 

 vance that to some extent resemblance between ancestors 

 and progeny is in some xvay connected with the chromosomes. 

 This admission relieves us of the necessity of an exhaustive 

 review of the evidence which necessitates the admission. 

 Though the evidence has practically all been brought out 

 during the last twenty-five years, it is large in quantity ^nd 

 widely scattered. Nearly all the semi-popular books, not 

 to speak of the many serials in technical biolog}^, present 

 some of it. We may therefore restrict ourselves to such 

 aspects as will seme our purpose from time to time. 



Throughout the vast range of living beings the rule like 

 'produces like holds sooner or later. I say sooner or later 

 because there are many exceptions were we to limit the state- 

 ment to parents and their immediate offspring. No indi- 



