(a,b) Average maps of CE rates of (a) the tissue (, dark blue) and of (b) cell divisions (, green), cell rearrangements (, magenta) and cell shape changes (, cyan) overlayed for comparison. Time averages were performed between 14 and 28 hAPF. Scale bars: 2 10-2 h-1. In this Figure (and Figure 4—figure supplement 1), values larger than the local biological variability are plotted in color while smaller ones are shown in grey; a local transparency is applied to weight the CE rate according to the number of cells and hemi-nota in each group of cells; black outline delineates the archetype hemi-notum; the midline is the top boundary; circles filled in yellow are archetype macrochaetae. Black rectangular boxes outline the four regions numbered 1 to 4 described in the text. (c) Projection of a cell process along the local axis of tissue elongation: cell process component. Here, is unitary and has the direction of the local tissue CE rate : its bar therefore indicates the local direction of tissue elongation (Equation 16). For each cell process measurement , its component along can be determined by its projection onto and is noted (Equation 17). It is expressed as a rate of change per unit time, i.e. in hour-1, and is represented as a circle whose area is proportional to the component. (Left) If a cell process CE rate (here orange bar) is rather parallel to (dark blue bar), it has a positive component on (orange empty circle). (Middle) If the bar is at 45° angle to , its component on is zero. (Right) If the CE rate bar is rather perpendicular to , its CE rate has a negative component on (orange full circle). Note that the component of tissue CE rate along itself, , is the amplitude of tissue CE rate , and is positive by construction. The additivity of cell process CE rates to (Equation 1) implies the additivity of these components to (Equation 2). Note also that these circles have a different meaning from those used to represent the dilation rates (Figure 1d). (d–g) Maps of components along the tissue CE rate for (d) the tissue rate itself (, dark blue, components are positive by construction) and of (e) cell divisions (, green), (f) cell rearrangements (, magenta) and (g) cell shape changes (, cyan). Same representation as in (a,b). Scale bars: 0.1 h-1. (h–k) Time evolution of components along the tissue CE rate of the tissue rate itself (, dark blue) and of cell divisions (, green), cell rearrangements (, magenta), cell shape changes (, cyan) and delaminations (, black) in four regions of interest. Measurements are averaged over sliding time windows of 2 h and spatially over the region (h) 1, (i) 2, (j) 3, (k) 4. In region 1, one can distinguish three phases where tissue morphogenesis mostly occurs via: (14–18 hAPF) cell rearrangements as divisions and cell shape changes nearly cancel out; (18–23 hAPF) cell shape changes, and it reaches its peak; (23–26 hAPF) oriented cell divisions as cell rearrangements and cell shape changes cancel out. In region 2, the same temporal phases can be distinguished: (14–18 hAPF) the effect of divisions is almost cancelled by cell shape changes; (18–22 hAPF) only weak changes occur; (22–26 hAPF) cell rearrangements and cell shape changes add up and morphogenesis becomes significant. In region 3, like in region 1, the two waves of oriented cell divisions can be seen clearly with peaks occurring around 16 and 23 hAPF, but here both division waves have a negative component along tissue CE rate. Cell rearrangements and cell shape changes make up for the negative sign of divisions, thereby mostly accounting for tissue morphogenesis in this region. In region 4, from about 18 hAPF onwards, the tissue CE rate significantly increases and almost perfectly overlaps with cell shape change CE rate, meaning that tissue morphogenesis solely occurs via cell shape changes.