562 THE FORMS OF TISSUES [ch. 



same way, save only that the initial dimple, instead of being 

 fortuitous, has its constant place, determined by the physico- 

 chemical heterogeneity of the embryo. We may even go one step 

 further, and see (or imagine we see) in the formation of the gastrula 

 a physico-chemical or physiological turning-point, the segmentation 

 cavity being due (as we have seen) to an inward flow, and a reversal 

 of the current leading to that shrinkage which produces the gastrula. 



Fig. 216. Effect of shrinkage on a globule of gelatine. 

 After E. Hatschek. 



A note on shrinkage 



We have dealt much with growth, but the fact is that negative 

 growth, or shrinkage, is also an important matter; and just as we 

 find a whole series of phenomena to be based on the extension or 

 expansion of bubbles, vesicles, etc., so there is another series, 

 physically ahke and mathematically identical, which depend on the 

 shrinkage of a soHd or semi-fluid mass. After all, growth and its 

 converse go hand in hand, and a special case of shrinkage is that 

 surface-tension to which all the Plateau configurations are due. 

 One clear case, the gastrula, we have touched on, and we have 

 discussed another which led to the stellate dodecahedra of the Rush. 



As a cube of gelatine, or of paraffin, dries, and shrinks, it alters 

 its shape in a remarkable way*. Its corners become more salient, 

 its sides become concave ; its cross-section has the form of a four- 



* Emil Hatschek. Kolloid Ztschr. xxxv, pp. 67-76, 1924; Nature, 1st Nov. 1924. 



