240 Priestley and Pearsall .» — Growth Studies. II. 
Robertson ( 14 ) has contributed some of the most carefully controlled 
quantitative data yet available for study. He has actively followed both 
lines of attack outlined above, and has prosecuted both the mathematical 
analysis of his curves and also the establishment of the analogy between the 
S curve of growth and the curve for an autocatalytic chemical reaction. 
His later work shows, however, interesting modifications of his original 
position. 
He begins with the assumption that the autocatalyst, in the successive 
S curves shown in the growth of certain animals, may be some part of the 
protoplasmic complex itself, nuclear in the firSt curve and cytoplasmic 
in subsequent stages ( 11 , p. 587). Later, he evidently considers that 
the catalyst must be sought for amongst definite chemical substances, 
probably provided by the glands responsible for internal secretions known 
to be associated with growth. This change of position is important, 
because such a definite catalyst would probably be responsible for modifying 
the rate of a definite reaction or type of reaction, and such reactions could 
only form single items in the complex machinery necessary for growth. 
Further, the experimental investigation of possible catalysts leads 
Robertson to a separate consideration of the two regions of different 
inflexion in the S curve, upon which the same catalyst appears to produce 
different effects. 
A consideration of this change of position appears to us of value as 
leading to certain preliminary conclusions that seem to be important. 
Because the quantitative data of growth obtained under uniform external 
conditions can be represented upon a single curve, it by no means follows 
that throughout the whole period of growth the same internal factors alone 
remain operative. It is probably significant that Robertson’s continued 
study of the problem leads him (1) to assume first a greater complexity of 
internal conditions and (2) to consider separately the different regions of his 
curves. 
In attempting to obtain further insight into the significance of the 
S growth-curve we have carried the process of analysis somewhat farther, 
and have found it fruitful to consider this curve as consisting of three 
distinct regions in which different internal factors are operative. Under 
these circumstances a mathematical expression for the whole curve seems 
to have no especial significance. 
We propose to consider as examples two cases of plant growth for 
which data are available — the growth of yeast in a limited supply of nutrient 
medium and the growth of roots from cuttings as recorded in the first paper 
of this series. No suggestion is made that the arguments advanced are of 
general application. Indeed, in the case of the similar curves obtained for 
the growth of individual leaves (Gregory ( 7 )), it is clear that other factors 
have there contributed to produce the same result. Probably every case 
