Composite bosons made of two bosonic constituents exhibit deviations from ideal bosonic behavior due to their substructure. This deviation is reflected by the normalization ratio of the quantum state of N composites. We find a set of saturable, efficiently evaluable bounds for this indicator, which quantifies the bosonic behavior of composites via the entanglement of their constituents. We predict an abrupt transition between ordinary and exaggerated bosonic behavior in a condensate of two-boson composites.PACS numbers: 05.30.Jp, 03.65.Ud Introduction. The (fermionic) bosonic behavior of any elementary or composite particle is ultimately implied by the spin-statistics theorem [1, 2], which can be derived under many different assumptions [3]. For composite bosons made of two fermions, the Pauli principle that acts on the constituents modifies the ideally expected bunching behavior [5][6][7], and changes the bosonic commutation relation [4]. The statistics of composites was recently re-considered from the perspective of quantum information [4]. Both, in the many-body properties of Bose-Einstein-Condensates (BECs) [8][9][10][11][12][13][14][15][16] and in dynamical processes [17][18][19][20], entanglement between two fermionic constituents turns out to be the crucial ingredient to ensure bosonic behavior [4].While for atoms and molecules, the impact of the Pauli principle that acts on the constituent electrons is typically small [9], the question of the effective compositional hierarchy and the impact of bosonic and fermionic effects on a higher level remains open, e.g., for molecules made of two bosonic atoms, and it is lively debated for α-particles in nuclear physics [21][22][23]. For a composite boson made of two bound bosonic constituents, no Pauliblocking jeopardizes the multiple occupation of singleparticle states. One could therefore expect such compound to simply inherit the bosonic nature of its own constituents. However, as we show below, the behavior of two-boson composites can heavily deviate from the ideal, because the single-particle states of the constituents tend to be unusually often multiply populated, leading to a super -bosonic compound. Although all matter is ultimately made of fermions, any high-level composite that is made of two bosonic constituents will face such super-bosonic effects.The quantitative indicator for bosonic features in the many-body theory of composites is the composite-boson normalization ratio χ N +1 /χ N [4,[8][9][10][11]. However, even when the two-boson wavefunction is known, the complexity of the algebraic expression for χ N +1 /χ N renders an evaluation for large N unfeasible [4].Here, we solve this problem by providing tight, saturable bounds for the normalization ratio, which allow us to efficiently characterize two-boson composites via three easily accessible quantities: the number of composites N , and the purity P and the largest eigenvalue λ 1 of the reduced density matrix of one constituent boson, which can be obtained from the two-boson wavefunction. This...