268 NATURAL SCIENCE. April. 



the idea is not so definitely formulated as in de Vries' chapter on 

 " the mutual independence of the hereditary characters," we are con- 

 stantly met with the assumption that the aspects of the organism, or 

 of its parts, looked at from a hundred different points of view, and 

 selected for consideration in a rich variety of ways, have a special 

 significance as units for the study of inheritance, being treated of as 

 independent of one another in heredity and variation, and often, for 

 that reason chiefly, being regarded as represented in the germ by 

 independent and special material vehicles. A dog, for instance, 

 according to the view of that writer, inherits some of his 

 characters from his mother, some from his father. Some come from 

 a remote ancestor, and others are quite new. One of the parents may 

 have had more to do with him than the other, for in the list of 

 characters it may be that there are ten from the father and only seven 

 from the mother. In most cases the characters are not the mean 

 between those of his parents, they do not even vary all in one 

 direction. There is no obvious interdependence between them. They 

 are, therefore, according to de Vries, to be referred to different material 

 foundations in the germ, derived from various ancestors, which 

 Anlagen may vary independently of one another, each working out the 

 character it represents. There are material vehicles for each of his 

 characters, as, among others, for the length of his legs, for his swift- 

 ness, for his bark, for his love of sugar and his fear of snakes, as well 

 as for the absence of all that he is not. All these are hereditary 

 characters, varying to all outward seeming independently of one 

 another, and all due to the ultimate Anlagen. Now, while few go so 

 far in isolating characters as de Vries does, yet their independence 

 of one another and their unity in themselves is so generally accepted 

 that it is worth while studying what is the principle on which 

 phenomena are grouped together and called a character, thereafter to 

 be treated of as unit, and on what principle they are grouped to form 

 two or more characters, varying independently or in correlation with 

 one another. 



The term may, in the first place, be used generally, of the whole 

 creature as individual, with or without any special reference as of 

 size, colour, habit, etc. ; and this is generally Weismann's use of the 

 term. Here there is no artificial division of the organism into parts 

 that have no answering physiological divisions. In agreement with 

 this point of view, Weismann's determining foundations in the germ 

 answer each of them to the complete character of one individual 

 ancestor, which is impressed on the whole of the new organism by 

 the little piece of matter in charge of that character and by the little 

 pieces derived from it, whereas those of de Vries answer each of them 

 to one artificially limited character of many ancestors, which aspect 

 of the new organism they look after, neglecting the rest of the body. 

 As Weismann often, however, uses the term character in its narrower 

 sense, he is led into the common fallacies at times. Indeed, an 



