Following Florys ideality hypothesis the chemical potential of a test chain of length $n$ immersed into a dense solution of chemically identical polymers of length distribution P(N) is extensive in $n$. We argue that an additional contribution $delta
mu_c(n) sim +1/rhosqrt{n}$ arises ($rho$ being the monomer density) for all $P(N)$ if $n ll <N>$ which can be traced back to the overall incompressibility of the solution leading to a long-range repulsion between monomers. Focusing on Flory distributed melts we obtain $delta mu_c(n) approx (1- 2 n/<N>) / rho sqrt{n}$ for $n ll <N>^2$, hence, $delta mu_c(n) approx - 1/rho sqrt{n}$ if $n$ is similar to the typical length of the bath $<N>$. Similar results are obtained for monodisperse solutions. Our perturbation calculations are checked numerically by analyzing the annealed length distribution P(N) of linear equilibrium polymers generated by Monte Carlo simulation of the bond-fluctuation model. As predicted we find, e.g., the non-exponentiality parameter $K_p equiv 1 - <N^>/p!<N>^p$ to decay as $K_p approx 1 / sqrt{<N>}$ for all moments $p$ of the distribution.
Angular correlations in dense solutions and melts of flexible polymer chains are investigated with respect to the distance $r$ between the bonds by comparing quantitative predictions of perturbation calculations with numerical data obtained by Monte
Carlo simulation of the bond-fluctuation model. We consider both monodisperse systems and grand-canonical (Flory-distributed) equilibrium polymers. Density effects are discussed as well as finite chain length corrections. The intrachain bond-bond correlation function $P(r)$ is shown to decay as $P(r) sim 1/r^3$ for $xi ll r ll r^*$ with $xi$ being the screening length of the density fluctuations and $r^* sim N^{1/3}$ a novel length scale increasing slowly with (mean) chain length $N$.
