The bilateral prototype (23) is inserted, and elastic impression material molded over it as it was over its bilateral counter-part — palatal surface first, then buccal. The impression material is extended anteriorly on both surfaces until it meets its bilateral counterpart.

A full arch plaster impression (24) is taken directly over the elastic impression material. This will pick up the uncemented processed partial bridge (25) with its cantilevered pontics, together with the elastic impression(s) (26, 27) and their embedded pterygoid extension implant prototypes (28, 29). The plaster pre-vents distortion of the elastic impression(s), and movement of the bridge within the soft impression(s).

On the stone model the dentist designs the implant, the posterior portion of which resembles that of the pterygoid extension implant for a totally edentulous arch : its buccal strut (30) — fenestrated for better tissue readaptation — wraps over the tuberosity distal to the sinus; its hamular notch portion (31) — the narrowest strut — is wedged against the pterygoid process; and its palatal strut (32) — the widest strut — extends anteriorly as a flat, relatively wide brace on the dense cortical bone where ridge and palatal bone blend. The modified Andrews or Dolder bar that will support the restoration and keep it from pressing on the crest arises from a strut (33) crossing the crest in the tuberosity region, behind the sinus. This cross-strut, called the tuber strut, is slightly wider than the hamular notch or pterygoid strut.

Anteriorly, of course, the implant differs. Instead of the fenestrated framework covering the premaxillae, the palatal arm expands into a fenestrated or solid "paddle." The reach of this paddle depends upon the length of the span and the strength of the anterior abutments. Typically in a short span with strong anterior abutments, the paddle extends a short distance (34). The longer the span and the less sturdy the anterior situation, the longer and

128