Bimetallic catalysts have long been known for their attractive catalytic properties, which are often superior to those of the components and which are generally found to depend strongly on the composition of the individual bimetallic nanoparticles making up the catalyst. [1][2][3][4] The improved catalytic performance had been explained in terms of concepts such as the electronic ligand effect or the geometric ensemble effect. [2,5] The fundamental understanding was hampered, however, by the largely unknown surface composition and, in particular, by the unknown distribution of the respective components in the surface layer. This led to a renewed interest in the chemical properties of the alloy surface where, in contrast to bimetallic nanoparticles, scanning tunneling microscopy (STM) makes it possible to directly evaluate the specific geometries available for adsorption on these surfaces, by high-resolution imaging with chemical contrast. Furthermore, density functional theory (DFT) calculations have opened up the possibility of studying the energetics of such sites. In this way, both ordered and disordered surface alloys, [6][7][8][9][10] overlayer structures, [11][12][13][14][15][16][17] steps, [18][19][20] and step modification [21] have been studied in some detail. Herein, we address the question of size effects in surface alloying and their impact on the chemical properties of the surface. We use a combined experimental and theoretical approach to quantitatively study how the surface chemistry depends on the size of islands of one metal in a matrix of another metal. The specific system we study is the adsorption of CO on well-defined bimetallic PdAu surface alloys. These surface alloys can be considered as model systems for PdAu catalysts, which have attracted considerable interest for several applications, most prominently for vinyl acetate synthesis. [22,23] The surface composition and distribution of surface atoms are determined by high-resolution STM. The interaction of CO with the respective surface is characterized by temperature-programmed desorption (TPD) and high-resolution electron energy-loss spectroscopy (HREELS). For the TPD measurements, it was verified that the surface was not modified by the temperature scan. The results allow us to clearly distinguish between electronic ligand effects, geometric ensemble effects, and adsorbate-adsorbate interactions. They also demonstrate the kind of agreement achievable between state-of-the-art theoretical work (DFT calculations) and experimental data, and the quality of chemical information extractable from high-resolution STM measurements in combination with statistical evaluation and nonlocal spectroscopic measurements. During the course of this work, Goodman and co-workers published a combined TPD/IR spectroscopy study on the interaction of CO with equilibrated, MoA C H T U N G T R E N N U N G (110)-supported PdAu alloy films, whose surface composition was determined by low-energy ion scattering. [23,24] While the distribution of surface atoms could not...