6oO REGENERATION AND GROWTH 7 



differentiation may begin as early as growth, though serial culture of cells in vitro 

 shows that it is held in check during rapid proliferation, possibly by a "dilution" 

 mechanism, like the kappa particles of Paramecium (Sonneborn, 1948). Cer- 

 tainly there is considerable overlap between the processes of growth and differen- 

 tiation, and the syncytial skeletal muscle-fibres of Vertebrates continue to grow and 

 proliferate nuclei long after typical differentiation is evident. A good deal of 

 macro- and micro-scopic differentiation in fact is simply differential growth. The 

 amphibian limb-regenerate grows in length much more than in width (Singer 

 and Craven, 1948) largely correlated with the elongation of the differentiating 

 muscles. At the molecular level there are three possible modes of differentiation, 

 by the rearrangement or by the chemical conversion of existing molecules or by 

 the intussusception of new types of molecule, — this last a growth-process. 



Visible differentiation in a regenerate begins at a greater absolute size in a large 

 than in a small individual (Paulain, 1938) and if the growth of a regenerate is re- 

 tarded then the regenerate may differentiate, normally, at an abnormally small size. 

 These facts indicate some independence between differentiation-rate and growth- 

 rate, but if retardation is severe then differentiation also is abnormal (Paulain, 

 1938; Needham, 1950a) and the abnormality has a quantal aspect i.e. com- 

 ponents either differentiate completely and normally or are completely lacking. 



The role of quantitative, rate-differences in controlling differentiation, has been 

 stressed by Child and his school (Child, 1941). The integrity and organisation of 

 a regenerate depends on a system of differential control, proportional to a rate- 

 system, differential in space. Brondsted (1955) finds support for this interpretation. 

 If the normally smooth gradient in rate from the mid-line to the lateral edges of a 

 planarian body-regenerate is interrupted by an appropriate procedure, then two 

 heads develop side by side at the anterad surface. Chandebois (1952) has thrown 

 further light on the nature of this "dislocation" between right and left sides but 

 her results essentially confirm the qviantitative factor in both normal and abnor- 

 mal differentiations. A piece cut from the lateral edge of a regenerate takes longer 

 to re-regenerate than a piece from near the mid-line (Brondsted, 1955), again 

 a quantitative manifestation. Chela-reversal during its regeneration in certain 

 Crustacea (Huxley, 1932; Darby, 1935, 1939) and similar compensatory rever- 

 sals (Abeloos, 1953; Ludwig and Ludwig, 1954) as well as the qualitative abnor- 

 malities known as heteromorphs (Przibram, 1934) all seem to be controlled by 

 essentially quantitative differentials. 



Most heteromorphs are qualitative abnormalities along the long axis of the 

 body. Some of these are relatively easily controlled by quantitative changes, for 

 instance in an intrinsic gradient of electrical potential (Barth, 1934; Marsh and 

 Beams, 1952; Moment, 1953). One weakness of Child's theory (Needham, 1952) 

 is • the stress on a single quantitative morphogenetic gradient. The electrical 

 (Moment, 1954) and other manifestations (Huxley and De Beer, 1934) of quanti- 

 tative gradients in the main axis of the annelid body in fact are amphiclinous. 

 If it be admitted further that these gradients must be much more complex in 

 local detail, and that some metabolic processes do, but others do not (Von 

 Bertalanffy, 1942), coincide with morphogenetic gradients, then the application 

 of the "quantitative theory" becomes reasonably clear. 



