Numbers of matrix- and analyte-derived ions and their sum in matrix-assisted laser desorption ionization (MALDI) of a peptide were measured using 2,5-dihydroxybenzoic acid (DHB) as matrix. As for MALDI with ${alpha}$-cyano-4-hydroxy cinnamic acid as matrix, the sum was independent of the peptide concentration in the solid sample, or was the same as that of pure DHB. This suggested that the matrix ion was the primary ion and that the peptide ion was generated by matrix-to-peptide proton transfer. Experimental ionization efficiencies of $10^{-5}-10^{-4}$ for peptides and $10^{-8}-10^{-7}$ for matrices are far smaller than $10^{-3}-10^{-1}$ for peptides and $10^{-5}-10^{-3}$ for matrices speculated by Hillenkamp and Karas. Number of gas-phase ions generated by MALDI was unaffected by laser wavelength or pulse energy. This suggests that the main role of photo-absorption in MALDI is not in generating ions via a multi-photon process but in ablating materials in a solid sample to the gas phase.