A strict interpretation of the size-principle hypothesis (37, 39-41) is that a muscle's motor units should be recruited in an ascending order according to both the size of their motoneurons and the size of their innervated muscle units (for reviews see Refs. 9, 39, 73). Studies of large mixed muscles in the cat hindlimb, however, have shown that motor axonal conduction velocity and tetanic tension, which are frequently considered indices of motoneuron and muscle-unit size, respectively, are uncorrelated in the fast-twitch (type F) motor-unit subpopulation (12, 13, 23, 24, 30, 32, 63, 71, 79). Attempting to focus on type F units, we compared the recruitment order of 42 pairs of cat plantaris (PL) motor units with both axonal conduction velocity and tetanic tension as well as with other muscle-unit properties. Single PL alpha-motor axons were functionally isolated in intact L7 ventral root filaments of decerebrate cats. Tension responses produced by stimulating each isolated motor axon were used to find the tetanic tension of the muscle unit and to classify the unit (12) as either type S (slow twitch, fatigue resistant), type FR (fast twitch, fatigue resistant), type FI (fast twitch, intermediate fatigability), or type FF (fast twitch, highly fatigable). Conduction velocity of each isolated axon was computed from measurements of axonal conduction time and length. The axon's reflex discharges were recorded from the proximal end of the cut filament and compared with the discharges of another PL axon residing in a different, previously cut filament of the same cat. The recruitment order of each motor-unit pair studied was found during reflexes elicited by homonymous muscle stretch, tendon taps, or single shocks at group I intensity to the PL nerve. If either axon of the pair failed to discharge, as often was the case with the high-threshold type F units, the monosynaptic reflex was facilitated by a 500-pps conditioning train applied proximal to a complete reversible cooling block of the PL nerve. In all 42 pairs studied, the weaker motor unit had the lower functional threshold for recruitment. Recruitment also invariably followed the order S greater than FR greater than FI greater than FF units. The motor unit of a pair with the higher resistance to fatigue thus always had the lower functional threshold. In 21 of the 22 pairs containing at least one type S motor unit, the unit with the slower-conducting motor axon had the lower functional threshold for recruitment.(ABSTRACT TRUNCATED AT 400 WORDS)