Measurements of the aerosol size distribution from 11 nm to 2.5 microns were made in Mexico City in March 2006, during the MILAGRO (Megacity Initiative: Local and Global Research Observations) field campaign. Observations at the urban supersite, referred to as T0, could often be characterized by morning conditions with high particle mass concentrations, low mixing heights, and highly correlated particle number and CO₂ concentrations, indicative that particle number is controlled by primary emissions. Average size-resolved and total number- and volume-based emission factors for combustion sources impacting T0 have been determined using a comparison of peak sizes in particle number and CO₂ concentration. Peaks are determined by subtracting the measured concentration from a calculated baseline concentration time series. The number emission and volume emission factors for particles from 11 nm to 494 nm are 1.56 × 10¹⁵ particles, and 9.48 × 10¹¹ cubic microns per kg of carbon, respectively. The uncertainty of the number emission factor is approximately plus or minus 50 %. The mode of the number emission factor was between 25 and 32 nm, while the mode of the volume factor was between 0.25 and 0.32 microns. These emission factors are reported as log normal model parameters and are compared with multiple emission factors from the literature. In Mexico City in the afternoon, the CO₂ concentration drops during ventilation of the polluted layer, and the coupling between CO₂ and particle number breaks down, especially during new particle formation events when particle number is no longer controlled by primary emissions. Using measurements of particle number and CO₂ taken aboard the NASA DC-8, the determined primary emission factor was applied to the Mexico City Metropolitan Area (MCMA) plume to quantify the degree of secondary particle formation in the plume; the primary emission factor accounts for less than 50 % of the total particle number and the surplus particle count is not correlated with photochemical age. Primary particle volume and number in the size range 0.1–2 μm are similarly too low to explain the observed volume distribution. Contrary to the case for number, the apparent secondary volume increases with photochemical age. The size distribution of the apparent increase, with a mode at ~250 nm, is reported.