Partially due to the absorption of the capping agents (stabilizers, etc) and also because electronic transition take place at shorter wavelength. What is your capping agent? Absorption below 250 nm is very common in organic compound.
The peak @390 nm in the absorption spectrum is the surface plasmon resonance band of silver and for gold they appear at longer wavelengths. As the relative percentage of Ag decreases in the nanoparticle the Au band starts becomes prominent and for pure gold nanoparticle the surface plasmon band is at ~520 nm.
The peaks @290 nm is not there in all the samples. It is there in only 3 samples (half the samples). It is not there in pure AgNP sample, pure AuNP sample and Ag:Au (50:25) sample but present in all other samples.
To answer your question, more details would be needed especially the experimental details such as preparation protocol, stabilizer used, solvent used etc.
Does your purification procedure differ from one sample to another?
I am assuming all the experiments were carried out under identical conditions. Just a note, you are changing the Ag:Au ratio in a very unusual way.
What do you mean by 'separated pyrex'? pyrex is just a material used to prepare glassware resistant to heat. What solvent do you use? It is well established that Au and Ag nanoparticles require some form of stabilizing agent, otherwise the particles aggregate to form larger particles and eventually the nanometric dimensions are lost. Do you use Ag and Au powders for a laser ablation process? It is possible that you are producing subnanometric particles or clusters along with nanometric particles. Very small Au and/or Ag particles do not exhibit plasmon bands but behaves similar to molecules and absorbs at shorter wavelengths, ex. Au11, Au22, Au39. You need to look for information about this. Could you describe your process you use to prepare your NPs? Do you store your nanoparticles under vacuum?
By the way, based on the profile of your absorption spectra you are also producing core-shell nanoparticles, for instance, the spectrum for AgAu 12.5:25 clearly shows two independent absorption bands, which is characteristic of core-shell structures. AuAg Alloys produce only one absorption band which shifts to lower wavelengths with increasing Ag content.