I note that this is an old question but thought it deserved an answer regardless.
The question is quite open ended and the easiest answer to give is that the sample is bimodal in nature and is therefore measured as such.
Reading between the lines, I might assume that you expect your sample to be monomodal and are surprised by the extra component.
The first thing to think about here is why you expect it to be monomodal?
Electron microscopy or other particle counting techniques may infer a monodispersed sample, but in fact only a small representation of the material is being observed and other components may be omitted by the sampling, preparation or even analysis.
If the sample IS truly monodisperse then the presence of extra size peaks may be a consequence of one of many factors-
Scattering artefacts such as scattering from the cuvette.
Non random motion such as sedimentation of thermal convection
Number fluctuations because of a sub-optimal sample concentration.
Different modes of diffusion, such as for non spherical particles where both rotational and transnational diffusion coefficients can be detected but both are interpreted as sizes by the analysis.
I could go on.
TLDR version, reassess why you think the sample should be monomodal and look at the quality of the DLS result. Without seeing any data or knowing anything about the sample, it is impossible to give an informed answer, which is probably why this has been unanswered thus far.