Because ILA was encountered in TCA-ethanol extracts of plasma dialyzed in boiled Visking casing which interfered with measurement of the "synalbumin insulin antagonist" on isolated muscle, a systematic study of the influence of the casing on these factors was undertaken. Plasma from healthy humans was extracted with TCA-ethanol and dialyzed in nonboiled and boiled Visking casing. The extract ("albumin") was tested by measurement of glucose uptake in rat hemidiaphragms, without and with insulin (1,000 /iU./ml.).. Significant amounts of both ILA and anti-insulin activity were detected in both "albumins" at 3 to 5 gm. per cent; however, more ILA and less insulin inhibition was found in "albumin" dialyzed in boiled casing. The magnitude of ILA could not be accounted for by immunoreactive insulin levels in the "albumin," and the ILA was not suppressed by insulin-specific antiserum (NSILA).From dose-response curves measuring glucose uptake and glycogen synthesis, more NSILA was detected in the "albumin" from boiled casing; however, at 3 to 5 gm. per cent concentration, "albumin" dialyzed in either casing caused inhibition of NSILA. Because partially purified NSILA was found to be retained equally by either boiled or nonboiled casing, an antagonist in Yisking casing was sought. A substance from Visking casing with small molecular weight was found to cause insulin inhibition. When nonantagonistic crystalline albumin was dialyzed in nonboiled casing, marked antagonism to both crystalline insulin and partially purified NSILA was demonstrated.It is concluded that an artifactual casing antagonist to both NSILA and insulin is bound to albumin and may contribute to the measurement of the "synalbumin insulin antagonist.