We compute CBR anisotropies in mixed models with different hot components , including neutrinos or volatile HDM arising from the decay of heavier particles .
The CBR power spectra of these models exhibit a higher doppler peak than CDM , and the discrepancy is even stronger in volatile models when the decay gives rise also to a neutral scalar .

CBR experiments , together with Large Scale Structure ( LSS ) data , are then used to constrain the space parameter of mixed models , when values of the primeval spectral index n > 1 are also considered .
Even if n > 1 is allowed , however , LSS alone prescribes that \Omega _ { h } \raise - 2.0 pt \hbox { \hbox to 0.0 pt { \hbox { $ \sim$ } } \raise 5.0 pt \hbox { $ < $% } } 0.30 .

LSS can be fitted by taking simultaneously a low derelativization redshift z _ { der } ( down to \simeq 600 ) and a high n , while CBR data from baloon–borne experiment cause a severe selection on this part of the parameter space .
In fact , while late derelativization and n > 1 have opposite effects on the fluctuation spectrum P ( k ) , they sum their action on the angular spectrum C _ { l } .
Henceforth n \raise - 4.0 pt \hbox { \hbox to 0.0 pt { \hbox { $ \sim$ } } \raise 5.0 pt \hbox { $ > $ } } 1.3 seems excluded by baloon–borne experiment outputs , while a good fit of almost all CBR and LSS data is found for \Omega _ { h } values between 0.11 and 0.16 , n \sim 1.1 and z _ { der } \sim 2000–5000 .
A smaller n is allowed , but z _ { der } should never be smaller than \simeq 1200 .