The work is motivated by the experimental observation of B. Chu [Macromolecules 28, 180 (1995)] which suggests that, while polymer globules in dilute solution in poor solvent are supposed to be very sticky, in actuality they collide many hundreds of times before merging and before aggregation starts. We argue that this slow-down is caused by an "entanglement force" which is operational on the prereptational time scale. This force arises from the fact that two touching globules cannot enjoy the mixing entropy gain expected in equilibrium until after they explore all conformations, including entangled ones. We report a molecular dynamics simulation in which we were able to measure the entanglement force as a function of distance. The important conclusion we can formulate so far is qualitative; the entanglement force exists and is sufficient to explain the observed slow-down of aggregation. (C) 2000 American Institute of Physics. [S0021-9606(00)50214-1].