Extended Data Fig. 2: Examples of the best-fit results of the Spergel profile fitting for our sample of galaxies on the uv-plane.

From left to right, we show the dirty image (natural weighting), source model convolved with the dirty beam, residuals after subtracting the source model and the normalized visibility amplitudes as a function of uv-distance. Contours start at ±3σ and increase by a factor of 1.5. The black crosses mark the centres of submm emission from sources derived from Spergel model fitting and the source name and the S/N_beam of the data are indicated in the left panels. The ALMA beam is shown in the bottom-left corner of the left panels. Solid red curves in the right panels represent the best fit of Spergel modelling to the uv-data. For comparison, an exponential (Sérsic n = 1) model is overlaid by a dashed black curve. Error bars show the statistical noise on the average amplitude in each bin. The reduced chi-squared values calculated for the best fit of Spergel modelling (\({\chi }_{\nu ,{\rm{Spergel}}}^{2}\)) and n = 1 profile fit (\({\chi }_{\nu ,n=1}^{2}\)), along with a degree of confidence (P(n = 1)) by which the n = 1 solution can be rejected, are reported in each panel.