A tandem spindle is defined a one in which several encapsulated sensory regions are linked together by some degree of muscle fibre continuity. Generally only one or two muscle fibres are continuous from capsule to capsule, while the rest taper off and partake of only one encapsulation. In the most common type, two capsules occur in linear succession (double tandem); triple and quintuple arrangements were also observed but not quadruple ones. Of 64 complete tandem spindles from cat rectus femoris cut in transverse section, 88% were double, 9% triple, and 3% quintuple. As measured in sectioned adult rectus femoris, the length of tandem spindles was 8.43 to 22.30 mm, (S. D.); by contrast, the length of 144 single spindles was 2.40 to 13.68 mm, mean 7.02 ± 2.45 mm (S. D). A study of the afferent innervation of 27 tandem spindles from the mesial flexor digitorum longus, semitendinosus, soleus, and rectus femoris muscles showed that the typical pattern was for a complex or intermediate capsule to be linked with one or two simple capsules. A quarter of the double tandems had complex or intermediate/intermediate capsule combinations; combinations in which all the capsules were simple, or all complex, did not occur. The primary ending in the simple capsules is invariably either holly or predominantly irregular, and the intramuscular diameter o f the fibre supplying it is usually less than that of fibres supplying other primary endings. On the basis of a study of serial transverse sections of a pair of kitten recti, and of twelve blocks sampled from the recti of six cats, it is concluded that 16 to 20% of the spindles in the cat’s rectus femoris are of the tandem type. Access to quantitative data gathered in connexion with other work suggests that, as compared with rectus femoris, the proportion of tandem spindles in semitendinosus and mesial flexor digitorum longus is slightly higher; in soleus, about the same; and in the pes interossei, considerably lower. An analysis of the results of teasing seven entire adult rectus femoris muscles indicates the following conclusions with regard to the proprioceptive equipment: ( a ) the intermediate type of single spindle occurs, on the average, with about double the frequency of either simple or complex types; ( b ) primary and secondary endings are present in about equal proportion in the muscle; ( c ) there are usually fewer tendon-organ endings than primary endings (average 80% ), but the proportion varies widely (41 to 139%) in different muscles; ( d ) paciniform corpuscles are relatively scarce and there appears to be no correlation between their frequency of occurrence as compared with other proprioceptors. The soleus, semitendinosus, and mesial flexor digitorum longus muscles appear to resemble rectus femoris with respect to ( a ) and ( b ), but in the fourth internal intercostal muscle complex spindles appear to be most frequent and secondary endings more abundant. The position and measurements of 144 single spindles in rectus femoris indicate that the lengths of the poles are related to the position of the spindle in the extrafusal muscle bundles and its orientation with respect to nerve supply. A re-appraisal of Ruffini’s classic paper (1898) on muscle-spindles suggests that in several respects the description was based on incomplete spindles cut in teasing gold chloride preparations. It is also evident that, owing to the methods employed, counts of muscle spindles made by most previous workers should probably be regarded as counts of spindle capsules. Failure to distinguish between single and tandem types of spindle also vitiates many of the measurements made of spindle length. Some discussion is offered on whether there is any functional significance in the tandem type of spindle and the proportional distribution of spindle types in a muscle.