Fig. 8. Model depicting changing levels of growth activators and inhibitors synthesized by disc cells in relation to increases in cell number. This model is a variation of a model proposed by Nijhout to explain how rising levels of growth inhibitors may determine disc size (Nijhout, 2003). In Nijhout's model, imaginal disc cells produce growth activators and inhibitors, and the inhibitors sequester or inactivate the activators. The level of each factor depends on the number of cells at the time and the influence of the inhibitor over the activator. At first activator levels are high, inhibitor levels low and disc growth is approximately exponential. When inhibitor levels become sufficiently high, growth begins to slow. Nijhout's model proposes that inhibitor activity remains high throughout slow growth. Accordingly, if pixie mutant cells were sensitive to the presence of inhibitor, then pixie mutant clone growth would be expected to worsen towards the end of larval life and during metamorphosis as growth slows. Instead, pixie mutant clone growth (survival) improves at the end of larval life (see text and Fig. 6E,F), and the intensity of cell death in pixie mutant and Minute discs decreases shortly before pupation (Fig. 4). Thus, we propose that the presence of activator is necessary for inhibitor synthesis. When inhibitor levels reach a threshold, this triggers the slowing down of growth and inactivation of the activator, which in turn leads to downregulation of inhibitor synthesis. Thus the pixie mutant clone phenotype is strong at the beginning of the slow phase, when inhibitor levels are high, but weak at the end of larval life when inhibitor levels are low, and growth is slow. The graphs presented here potentially apply to both the hinge and the pouch, with higher levels of inhibitors being found in the pouch than the hinge.