To understand the mechanism of dynamic instability of microtubule growth and shortening, one needs a means of reliably determining the polarity of the microtubules under investigation. Sea urchin sperm-tail axonemal fragments nucleate the growth of both plus-ended and minus-ended microtubules, but their polarity is not apparent by video-enhanced DIC microscopy. The polarity of a microtubule is usually assessed by observing differences between the rates and lengths of growth and shortening excursions of the two ends. In practice, though, a significant fraction of the population of microtubules displays characteristics intermediate between the average characteristics of either end, thereby escaping classification. Excluding these "intermediate" microtubules from the measured populations introduces bias into the understanding of microtubule dynamic instability. We circumvent this problem by making use of the plus-end directed movement of the microtubule-dependent molecular motor kinesin to determine the polarity of any given microtubule unambiguously. Carboxylated-microspheres coated with kinesin, which are clearly visible by DIC microscopy, were used to determine the polarity of a microtubule. The dynamics were then observed. Kinesin was found to have no marked effect on dynamic instability. By this technique, we show that the distributions of properties that describe microtubule dynamic instability (rates and lengths of growth and shortening as well as frequencies of interconversion between these phases) of plus-ends overlap to a significant extent with those of minus-ends. It is this overlap that obscures the usual classification of the ends. Therefore, models describing microtubule dynamic instability need to incorporate the broad and overlapping range of properties of the two ends.