100 THE CONTROL OF LIFE 



do not read or strain their eyes ; it occurs in horses ; it 

 is probably due to a germinal variation. It seems that a 

 change impressed on a unicellular organism may be 

 handed on, but the distinction between * body ' and 

 * germ-cells ' has not been established at this low level. 

 (4) Reappearance in successive generations does not 

 prove transmission, unless the possibility of the modifica- 

 tion being hammered on afresh be excluded. (5) What 

 is exhibited at birth in a mammalian offspring or at 

 separation in a flowering plant's seed is not necessarily 

 part of the inheritance, it may have been impressed 

 on the offspring by the mother, through the medium 

 of the blood or the sap. The secondary or indirect 

 results of a maternal modification may thus affect the 

 unborn mammal or the unliberated seed, since both live 

 in close union with the parent. (6) Poisoning of the 

 whole system of the parent may cause deterioration of 

 the germ-cells, and yet not result in the transmission 

 of any particular modification. (7) Sheep transported 

 to a colder country may show after some months a change 

 in the character and length of the wool ; their progeny 

 may have even thicker and longer fleece. Is this not 

 proof enough of the transmission of a modification ? 

 The answer must be " Certainly not." For the second 

 generation were subjected to the peculiar nurture from 

 birth, their parents only from the date of transportation. 

 If there were measurable increase in the third generation 

 as compared with the second, and if no selection was in 

 process, there would be evidence worth considering. 

 But we do not know of its being forthcoming. 



