On the dispersion of imaginal progenitor cells in theDrosophilablastoderm

Abstract

SUMMARYThe number of blastoderm cells inDrosophilawhose descendants form adult structures has frequently been estimated from genetic mosaics. Data from somatic recombination (method I) and gynandromorph (method II) mosaics both yield very low estimates, e.g. about 10–20 progenitor cells for the eye and antenna, wing or leg.In gynandromorphs the mosaic dividing line has a random orientation on the blastoderm. In the 6000 cell blastoderm it should be very unlikely that the mosaic dividing line passes through any small patch of only 10–20 cells. Yet it has been reported that 10–25% of eye/antenna, wing or leg disks in gynandromorphs are mosaic. Thus the frequency of mosaicism data seems to be in contradiction to the progenitor population estimates. Similar discrepancies are found in the data for other adult structures.In this paper we derive a formula for estimating the number of cells in a blastoderm patch from the frequency with which the gynandromorph dividing line passes through it (method III). In a second method (method IV) we use the maximum distances inside the progenitor areas on a fate map to estimate the progenitor patch size. These two estimates agree closely with each other. We find, e.g. that 50–100 cells are in the patches from which the eye/antenna, wing or leg disks derive.We examine a number of possible explanations for why the first two estimates are so much smaller than the last two. The former estimates refer to the number of progenitor cells which actually have descendants in the adult structure; the latter estimates refer to the total patch area in which the progenitor cells sit. With the present information the most reasonable conclusion is that the progenitor cells for the adult structures are dispersed among other cells which have different developmental fates. If confirmed by experiment, this result has many implications for the process of determination.