Mankiw & Whinston (1986) show that the market may contain too many firms. The intuition is that if a firm charges a suboptimal high price, a rival has a demand curve that is "too high" and can also enter the market. This is socially suboptimal if both firms incur fixed costs and have the same marginal costs, because one firm could as well supply the total quantity alone with the same variable costs but lower total fixed costs.
What I don't understand is: Why are fixed costs needed for this resultat? Fixed costs might as well be absent. Then a suboptimal high price might allow a rival with higher marginal costs to enter the market. Note that the entrant might not even produce the same product but an imperfect substitute.
If possible, a graphical representation of the answer would be much appreciated.