DEVELOPMENT AND HEREDITY 1359 



growth of cancer-cells. In mice a cancer may be transferred from one 

 individual to another by inoculation, and this process may apparently 

 go on indefinitely, so that finally a mass of cancer-cells may have 

 been produced equal in volume to many thousands of mice, and per- 

 sisting long after the mouse from which it was first taken would have 

 died under natural conditions. 



In sexual reproduction the new individual partakes of charac- 

 teristics of both its parents. It therefore resembles neither of its 

 parents in all details. The conjugation of the two parent cells from 

 which it is derived has been preceded by a throwing out of half the 

 chromosomes from each parent cell. It is therefore natural to ascribe 

 the variations which occur among the members of one family to a 

 qualitative difference in the chromosomes which have been eliminated 

 in the formation of their respective egg-cells. Can we regard the 

 chromosomes as representing separate qualities of the individual, or 

 must we assume that all qualities are represented to a greater or less 

 extent in every chromosome ? In the case of many qualities, especially 

 those which distinguish the species as apart from the individual varia- 

 tion or family characteristic, we must probably accept the latter idea 

 as correct. In this case the child can be regarded as representing an 

 arithmetical mean of both its parents. In certain respects, however, 

 a quality seems to be transmitted from parent to offspring either com- 

 pletely or not at all. This is specially applicable to those characteristics 

 which have been rapidly produced by artificial selection, characters 

 which, if artificial selection be abandoned, rapidly disappear, with 

 reversion to the type from which the special strain was ultimately 

 produced. The way in which these characteristics are transmitted 

 was first studied by Mendel and has been formulated as Mendel's law. 

 Mendel's first experiments were carried out on peas. On crossing a tall 

 plant with a dwarf plant seeds were obtained from which all the plants 

 were tall. On recrossing the plants of this generation among one 

 another a third generation was obtained in which 25 per cent, of 

 the plants were dwarfs and 75 per cent, were tall. Crossing the dwarf 

 plants among themselves led to the production of dwarf plants 

 through successive generations. Of the tall -plants 25 per cent, arid 

 all their descendants continued to produce tall plants when self- 

 fertilised, whereas of the remaining 50 per cent, of the tall plants 

 25 per cent, produced dwarfs and the remaining 75 per cent, produced 

 tall plants. On continuing the process of breeding, the dwarf plants 

 when self -fertilised always produced dwarfs, whereas of the tall plants 

 25 per cent, produced tall plants, which bred true, while the remain- 

 ing 50 per cent . produced the same percentage of tall and dwarf as in 

 the preceding generations. Mendel explained these results by the 

 assumption that a character could be dominant or recessive. If both 



