Differentiation and Morphogenesis in Insects 253 



recall that the metamorphosis of the animal as a whole is a mosaic of 

 metamorphoses at the level of the individual cells. And here we come 

 face-to-face with the most baffling phenomenon of all. It turns out that 

 the individual cells at an early stage in emhryonic development are 

 "determined" or "programmed"" for the specific sequence of differ- 

 entiations which days, weeks, or months later they will act out in the 

 future larva, pupa, and adult. In this sense, the construction manual is 

 more akin to the script of a three-act play in which the individual 

 cells are cast in individual roles. 



Metamorphosis of skin implants 



The programming of cells is particularly well illustrated in the ex- 

 periments first performed by Kiihn and Piepho ( 1938, 1940 1 and 

 subsequently continued in great detail by Piepho and his students at 

 Gottingen (Piepho, 1938a, b; 1943 ) . 



If a tiny fragment of skin is cut from a caterpillar and implanted 

 into another caterpillar of the same or even of a different species, the 

 implant not only survives but undergoes a complicated career of 

 growth and metamorphosis in synchrony with that of the host. 



The first thing that happens is that the epithelium of the implant 

 grows around to close on itself to form a cyst — a hollow ball with 

 walls only one cell thick. Curiously enough, this growing around 

 occurs in such a manner that the cuticular surface of the epithelium 

 always faces inward to form a lining to the cyst. 



Days or weeks later, when the host larva undergoes a larval molt, 

 so does the little hollow ball of cells. The cuticular lining is shed 

 into the lumen and each cell now secretes a new larval cuticle. This 

 shedding and re-forming of larval cuticle is repeated over and over 

 again as long as the host is undergoing larval molts. If the implant 

 is removed prior to the pupation of the host and re-implanted into a 

 sequence of immature larvae, there is apparently no limit to the num- 

 bers of larval molts that it can undergo (Piepho, 1943; Piepho and 

 Meyer, 1951 ) . So, as far as its cells are concerned, the larval molts 

 correspond to the use and re-use of the coded instructions for syn- 

 thesizing and secreting a larval cuticle. 



But, in the normal course of events, sooner or later the host meta- 

 morphoses to a pupa. So does the implant. The larval cuticle is cast 

 off into the lumen of the cyst and each cell now secretes an overlying 

 island of pupal cuticle. The cyst, in short, has pupated. Manifestly, 

 each cell is now tapping information that bad previously been re- 

 pressed, presumably within its own chromosomes. Still later, when 



