(a) Timing of the sequence of morphogenetic events at gastrulation. (b) Stretching of posterior endoderm cells by mesoderm invagination in the WT, followed by compression by GBE. Such stretching is shown to be absent in halo sna mutants. Representative of the N=6 embryos observed in each case. (c) Mechanical rescue of posterior endoderm stretching by highly localized magnetic interactions with UML-loaded mesoderm in halo sna mutants. (d) Chemographs of the posterior endoderm of WT, halo sna and magnetically stretched halo sna. Representative of the N=6 embryos observed in each case. (e) Distribution of Myo-II-GFP in sqh-GFP, halo sna sqh-GFP, halo sna sqh-GFP magnetically stretched posterior endoderm, and Fog RNAi-injected halo sna sqh-GFP magnetically stretched posterior endoderm. Representative of the N=6 WT embryos, the N=6 halo sna embryos P=2 × 10−3, exact Fisher test), the N=14 halo sna stretched embryos (P<10−3) and the N=7 halo sna Fog RNAi-stimulated embryos (P<10−3). Scale bars, 5 μm. (f) Invagination (red arrow) in the posterior endoderm of non Rnai-injected (representative of the N=6 embryos observed), sna, bcd, nos, tor Rnai-injected (representative of the N=4 embryos observed, P=2 × 10−3), sna, bcd, nos, tor Rnai magnetically stretched posterior endoderm (representative of N=5/8 embryos, P=0.02), and bcd, nos, tor Rnai embryos (N=4, P=0.03). Scale bars, 10 μm.