444 HEREDITY AND DEVELOPMENT 



remains latent until special circumstances arise which call it into 

 activity. 



Note on Regeneration. — When half of a highly differen- 

 tiated Infusorian like Stentor regenerates the missing half, we 

 suppose that it does so because in each half there are diffusely 

 distributed " specific units " or " groups of determinants," 

 which can in appropriate environment grow into wholes. We 

 are encouraged to hold this hypothesis since we know that slices 

 of Stentor a millimetre or less in thickness can re-grow wholes. 



We shift the experiment to a slightly higher level, and we find 

 that fragments of relatively simple multicellular animals, such 

 as Hydra and Planarians, can grow into entire organisms. We 

 suppose that the excised groups of cells have among them a 

 sufficient complement of " specific units " to ensure the de- 

 velopment of a complete organism. 



But as we ascend higher in the scale, we find that while 

 the earthworm can re-grow a new head or a new tail, a few 

 median segments cut out of the middle of an earthworm will 

 soon die. A crab can re-grow a lost limb, but the limb cannot 

 re-grow a crab. The inference is that as differentiation increases 

 the diffuse distribution of " complete specific units " ceases, so 

 that the excised part is no longer a viable fragment. All this 

 points to the reality of differential cell-division. 



If the eye-bearing horn of a snail be cut off, it is regenerated 

 over and over again, with the complex eye complete. If the eye 

 of a crab be excised, there is usually regenerated an antenna in- 

 stead of an eye, but if the optic ganglion is not injured a normal 

 eye is regenerated. If the front of the eye of a newt or of a 

 salamander be cut off, a new lens is regenerated. All this points 

 to the hypothesis that within limits, probably punctuated by 

 natural selection, the maimed stump or foundation of an im- 

 portant organ retains in reserve a contingent of units capable of 

 growing the whole of that organ. Thus, while the distribution of 

 complete residual specific units or ids becomes more and more 



