The expression of myosin heavy chain isoforms in muscle spindle fibres has been the subject of a number of immunocytochemical studies, some of them with discordant results. In order to assess whether these discrepancies are due to differences in the specificity and sensitivity of the antibodies used, we have compared the reactivity of rat muscle spindle fibres to two pairs of antibodies presumed to be directed against slow tonic (ALD 19 and ALD 58) and neonatal (NN5) and neonatal/fast (MF30) myosin heavy chains. Adult, developing and neonatally de-efferented muscle spindles from the rat hind limb muscles were studied in serial cross-sections processed for the peroxidase-antiperoxidase method. Important differences in the staining profiles of intrafusal fibres were noted when ALD 19 and ALD 58 were compared. ALD 19 stained the muscle spindle precursors from the seventeenth day in utero, whereas ALD 58 only did so by the twentieth day of gestation. In adult spindles ALD 19 stained the nuclear bag1 fibres along their entire length, whereas ALD 58 did not stain these fibres towards their ends. ALD 19 stained the nuclear bag2 fibres along the A, B and inner C region, but ALD 58 stained these fibres only in the A and the inner B regions. ALD 19 stained some nuclear chain fibres along a short equatorial segment, whereas ALD 58 did not stain the nuclear chain fibres at all. NN5 stained the nascent nuclear bag1 and chain fibre precursors at earlier stages of development than MF30. Clear differential staining between primary and secondary generation of both extra- and intrafusal myotubes was seen with NN5, whereas MF30 stained all myotubes alike. However, in postnatal spindles, MF30 was a very good negative marker of nuclear bag1 fibres. The staining profile of the adult fibres with NN5 and MF30 was rather similar. The staining pattern of neonatally de-efferented bag fibres obtained with ALD 19 and ALD 58 was practically identical and it differed from that of control spindles, confirming that motor innervation participates in the regulation of the expression of slow tonic MHC along the length of the nuclear bag2 fibres, as we have previously shown with ALD 19. The distinct staining patterns obtained with ALD 19 versus ALD 58 and with NN5 versus MF30 reflect differences in antibody sensitivity and specificity. These differences account, in part, for the discrepancies in the results of previous studies on muscle spindles, published by Kucera and Walro using ALD 58 and MF30, and by us using ALD 19 and NN5.