Abstract. We argue that nuclei are not transparent for fast projectile partons. Color transparency is effective for final state interactions though nuclear filtering of initial partons can suppress the hadroproduction in nuclei. We challenge the standard scenario of hadroproduction and obtain a qualitative agreement with the data. PACS: 24.85.+p; 13.85.QkThe data for hadroproduction on nuclei contradict the standard model of color transparency (CT): the suppression of J/~p and Y production in nuclei [i], as well as the large distortion of pr-distributions in all hadroproduction reactions, is far beyond the predictions of CT model. We argue that the hadroproduction suppression in nuclei due to hard initial state (IS) interactions of projectiles is not ruled out by the CT forecasts.The propagation of a color singlet partonic configuration in nuclei is described by cross sections vanishing like r 2, where r is its transverse size [2]. Heavy quarkonia are first produced as bare singlet quark-antiquark configurations with r,,~ 1/Q and further hadronize and get a normal hadronic size far outside the nucleus. Therefore the final state interactions for heavy quarkonium production in nuclei are unimportant. The situation with IS interactions is quite different. Many authors argued that only small size projectile parton configurations are involved in the annihilation with large Q2 (see, e.g., [3]). In fact, the annihilating (and produced) partons must have the transverse separation ~ 1/Q. But the annihilating partons belong to different hadrons (projectile and target). There is no restriction for transverse separation between quarks from the same projectile hadron. Therefore large size partonic configurations of projectile hadrons could contribute to hadroproduction as well. These Several authors [4][5][6][7] demonstrated the cancellation of IS gluon exchange diagrams in hadroproduction on nuclei at large Q2. We now focus on the contribution of hard active-spectator IS interactions in nuclei with exchanged momenta l,,-~ P, P is the lab. frame projectile momentum. Assume that the operators Q1 and Q2 describe the IS interactions and the annihilation of a projectile parton, resp., li) is its asymptotic initial state and If) is the final state after the interaction Q2. The vanishing of Feynman diagrams describing the two-step process, i.e. the annihilation after the IS scattering, means that (fl Q2 QIli) =0.( 1) Many authors concluded that the interaction Q1 can be neglected if the above equation is valid. This would be correct if the fields were infinite (remember that Feynman diagrams describe interactions of infinite fields). In this case we could use the same assymptotic initial state [i) for both interactions Q1 and Q2. In reality, particles are localized objects and asymptotic initial states for front and back nucleons can be different. We assume that the interaction Q1 is sufficiently short-ranged,L is the target size in the lab. frame. In this case the second step of the process is described by the amplitude (fl Qzli')...