Abstract. A method for proving the embeddability of semigroup amalgams is introduced. After providing necessary and sufficient conditions in terms of representations for the weak embeddability of a semigroup amalgam, it successfully deals with the embedding of inverse semigroup amalgams into inverse semigroups and the embedding of an amalgam of regular semigroups whose core is full in each member.Most of the known results on amalgamation of semigroups, almost all of which are due to J. M. Howie, have now been proved by a method introduced in [7] involving a countably infinite number of steps, each step extending a representation of a semigroup [7], [8], [14], [15]; the main results thus proved concern amalgamation over a common unitary subsemigroup, an almost unitary subsemigroup, an inverse subsemigroup, a two-element subsemigroup, and the embedding of an inverse semigroup amalgam in an inverse semigroup. We give here a method of proof that avoids this infinite number of steps.First it yields a necessary and sufficient condition, in terms of representations, for the weak embeddability of a semigroup amalgam (defined below).This gives then a short proof of [7, Theorem 8] (see also Howie's text [9] for an exposition), namely that inverse semigroups have the strong amalgamation property. Further we are able to show that finitehess can be preserved in the embedding of an amalgam (S¡, i E I; U) of inverse semigroups if the common inverse subsemigroup U is full in each S¡, i.e. contains all the idempotents of each S¡; in general, finiteness cannot be preserved [7, §3].Further we are able to show that, in fact, regular semigroups can be amalgamated over a full regular subsemigroup, also with finiteness being preserved. In general, an amalgam of regular semigroups (even left regular bands) cannot be weakly embedded in a semigroup [8, Remark 7]. In the final section we consider semigroups that are unions of groups.