|Select image to enlarge|
Origin of discs as islands within the embryonic ectoderm.
Left side of an embryo (below) at the cellular blastoderm stage (D, dorsal; V, ventral; A, anterior; P, posterior). Spots indicate where discs (key at right) or histoblast nests later form. In this fate map, segments (straight or warped rectangles) are numbered in each region (Head, Thorax, Abdomen). Areas shaded medium (dEpi, dorsal epidermis) or dark (vNR, ventral neurogenic region) form larval skin, except that ~25% of vNR cells ingress as neuroblasts. The pseudocephalon (H) is the only part of the head that remains superficial after gastrulation , which also internalizes the midgut (amg, pmg), mesoderm (mes), and pole cells (pc). Amnioserosa (as) is extraembryonic.
This fate map is based on Hartenstein's atlas  with the following exceptions: (1) D midline is not tilted toward viewer; (2) clypeolabral, eye, labial, and humeral sites are as per gynandromorph maps [1695, 1777, 2026, 4146] though the actual humeral site may be more ventral since its spiracle and the 1st-leg disc can share blastoderm cell ancestors [237, 928, 2819, 4565]; (3) eye anlage is shown as one oval instead of the three spots that Volker Hartenstein used (pers. comm.) as an arbitrary way to connote its mysteriously diffuse origin [1147, 1224, 2103] and its patchy apoptosis [3058, 4825]; (4) thoracic discs are more dorsal as per cell transplant data ; (5) wing and 2nd-leg disc are fused as per histological [827, 834] and lineage data [2819, 4076] as are haltere and 3rd-leg disc; and (6) genital disc spans A8-A10 [429, 735, 2343].
N.B.: H1-H3 are stylized  and probably not contiguous [832, 1300, 1631, 2104, 4825]. See [2103, 3058, 3631] for head and tail details and [631, 3791, 3792] for other nuances. In the schematic drawing of the larva (above) discs are spread out, and the midsection is omitted. See also App. 7.