304 



Embryogenesis: Progressive Differentiation 



It has been emphasized for the chick that 

 invagination in the primitive streak must 

 involve a de-epithelization process (Wadd- 

 ington and Taylor, '37) where individual 

 cells lose connection with other cells (detach 

 from the surface coat?) and become rounded 

 and heaped up in the streak. Evidence from 

 a wide variety of histological and histo- 

 chemical studies is continually accumulating 

 to indicate that this invagination process 

 is accompanied by a whole series of cellular 



:^7-v. 



Fig. 114. Healing of trout blastodisc after removal 

 of one-half its substance. Left side, morula stage. 

 The lower figure shows that wound healing and 

 closure take place evenly from all sides. Right, 

 gastrula stage. Only the ventral (top) half stretches 

 to heal the wound (see lower figure). (After Devil- 

 lers, '48a, Fig. 55.) 



changes: increased affinity for dyes in the 

 upper layei's of the streak at least; increased 

 reducing capacity for oxidation-reduction in- 

 dicators (Rulon, '35); loss of lipids upon 

 invagination (Jacobson, '38) ; increased ribo- 

 nucleic acid content followed by its loss in 

 laterally migrating mesoderm — not, evi- 

 dently, in the axial mesoderm which mi- 

 grates anteriorly (Gallera and Oprecht, '48) ; 

 increased sulfhydryl (reported by Brachet; 

 cf. also Burio, '51); concentration of indo- 

 phenol oxidase and phosphatase (Moog, '43, 

 '44). The anteroposterior gradient of activity 

 observed so frequently in these studies on 

 the streak is of course morphologically a 

 dorsoventral gradient of the blastopore 

 and may be at least in part a function of 

 differential massing and tempo of invagina- 

 tion of cells in the different parts of the 

 streak. 



In the teleost, invagination is primarily an 

 ingression of a continuous sheet of cells; 

 Devillers ('51b) has figured in Salmo the 

 process of loss of connection with the surface 

 coat, as in the amphibian gastrula, and has 

 also described loss of basophilia during the 

 process. According to Oppenheimer ('36b), 

 some individual detachment of cells from 

 anterior portions of the embryonic shield 

 epiblast, to join the mesoderm, occurs in 

 addition to ingression at the dorsal lip. It 

 is difficult to avoid comparing these de- 

 epithelizations to the more radical experi- 

 mental situation demonstrated by Holtfreter 

 ('47) in explants of amphibian ectoderm, 

 where neural differentiation appears to be 

 associated with the margin of the epithelial 

 mass, under conditions promoting slight 

 cytolysis and eventual sloughing of cells. 



The experimental analysis of the am- 

 phibian egg has led to the identification of 

 the region actively concerned in induction 

 of central nervous system, with the in- 

 vaginating chordamesoderm. It would be 

 expected from the distribution of the latter 

 areas that Salmo and Fundulus (cf. Fig. 

 112) might also differ in spatial extent of 

 material capable of performing the primary 

 induction. This property has not been com- 

 pared directly in the two forms. It is known, 

 for both, that a bit of dorsal lip or already 

 invaginated material will induce a super- 

 numerary axis from material that would not 

 normally form dorsal axis. A phenomenon 

 that would seem to bear some relation to 

 such direct tests of induction is the differ- 

 entiation of axial tissues in isolated pieces 

 or grafts taken from regions which normally 

 do not contribute to the dorsal axis. In 

 Salmo in the blastula stage, according to 

 Luther ('36a), all quadrants of the blastoderm 

 are equipotent for the formation of axial 

 structures when grafted to the yolk sac of 

 hatched fry. It is hard to imagine that the 

 neural tubes in these grafts arose by inde- 

 pendent differentiation of ventral ectoderm: 

 we are forced to interpret such formations as 

 "self-organizations" involving release, among 

 other things, of some latent neural-plate- 

 inducing potency somewhere in the isolate. 



The ability of the ventral side of the trout 

 blastoderm to form an axis when released 

 from the influence of the dorsal side was 

 confirmed (Luther, '37) by combining two 

 ventral half-blastulae: an axis regularly 

 formed, usually from the half that was slight- 

 ly younger. Devillers ('51a) in extending this 

 work has made the striking discovery that 

 if the preparation is made using as a host 



